Design of a Quadcopter for Winning the Jerry Sanders ...



Design of a Quadcopter for Winning the Jerry Sanders Creative Design CompetitionJordan Bramble, Harold Merida, Ibtsam KhanAurora Flight SciencesContents TOC \o "1-3" \h \z \u 1.0 Description of Goals of Competition PAGEREF _Toc373064268 \h 32.0 Competition Rules PAGEREF _Toc373064269 \h 83.0 Experiment PAGEREF _Toc373064270 \h 94.0 Strategy Analysis PAGEREF _Toc373064271 \h 175.0 Mission Requirements PAGEREF _Toc373064272 \h 256.0 Functional Requirements PAGEREF _Toc373064273 \h 267.0 Design PAGEREF _Toc373064274 \h 267.1 Design Requirements PAGEREF _Toc373064275 \h 287.2 Alternatives PAGEREF _Toc373064276 \h 287.3 Weighted Values PAGEREF _Toc373064277 \h 297.4 Parts List PAGEREF _Toc373064278 \h 317.5 Pick Up Mechanism PAGEREF _Toc373064279 \h 358.0 Simulation Design PAGEREF _Toc373064280 \h 379.0 Project Management PAGEREF _Toc373064281 \h 399.1 Schedule of Work PAGEREF _Toc373064282 \h 399.2 Budget PAGEREF _Toc373064283 \h 419.3 Project and Risk Mitigation PAGEREF _Toc373064284 \h 449.4 Work Breakdown Structure PAGEREF _Toc373064285 \h 4710.0 References PAGEREF _Toc373064286 \h 481.0 Description of CompetitionThe Jerry Sanders Creative Design Competition, sponsored by Advanced Micro Devices, and named after their former CEO, is a yearly competition in robotics held at the University of Illinois Urbana-Champaign. The 2014 competition will be held March 14th – 15th. Multidisciplinary teams of engineering students nationwide, participate in a game like competition for two days with robots they’ve designed and constructed. Each year the competition challenges change and require teams to develop innovative mechanisms for successfully scoring points. The competition takes place in a 2000 sq. ft. arena (44.7 x 44.7 ft). For airborne robots, a sloping net is hung above the course that is always approximately 6 ft above the arena floor.The goal of the competition is fostering innovation and creativity within robotics and promoting the engineering disciplines. Each year thousands of spectators attend. In last year’s competition, 26 teams from six different universities participated in the competition, attempting to pick up cones and place them on pins in order to score points. All of the participating teams hail from universities within Illinois or Indiana. Historically only two airborne entries have competed. During last year’s competition, the Northern Illinois University Robotics club entered a quadcopter and they were successfully able to pick up cones, however they could not accurately place them on pins and the downward thrust from the propellers would occasionally blow away unsecured cones before they could pick them up. Figure SEQ Figure \* ARABIC 1: An eagle eye view of the course The colored boxes in the corners correspond to each team’s starting area; the four colored rectangles represent ramps that lead up to the second level territories. At the base of the second level are small ‘soccer’ balls that are assigned by color to each team, these balls can be moved to their respective places, around the corner of the second level, in order to change direction of the ramps. The ramps are equipped with conveyor belts initially moving downward, but changing their direction causes the conveyor belt to change direction thus pulling a robot up. The central black box represents the third level territory, and the surrounding 8 boxes represent the second level. All of the other light blue dots represent the pins in the first level. The light purple boxes on each edge are designed to hold cones for each team. Figure SEQ Figure \* ARABIC 2: Hinged Door LocationThe competition shall consist of picking up and transporting cones of a specific color to pins corresponding to territories, where each pin is 1 inch in diameter. This is equivalent to controlling a “territory”, and the length of time a territory is controlled determines how many points are awarded. Before going into detail about how points are awarded it is necessary to understand where the territories are placed. Figure 3 is a CAD diagram used to measure the distances between territories. The type of territory is color coded.Figure SEQ Figure \* ARABIC 3: Cad Diagram of all the 3 Different TerritoriesGreen represents first level territories, of which there are 36. Yellow represents the second level, of which there are 8. Finally, there is 1 third level territory marked in red. It is also necessary to understand how cones are placed in the course.Figure SEQ Figure \* ARABIC 4: Pin Location in the ArenaFor the team starting in the green corner, their cones are arranged in this way. Ten cones are unstacked behind a drop wall that can be opened by pulling out two pins that hold the wall in place. Five cones are in front of their home territory, hanging from strings. Five more cones are placed behind a hinged door beneath the second level that must be pulled open. They also have a stack of five cones in their home territory as well as 5 cones stacked at the second level and five opposite of the raised platform. Since the course layout is symmetrical, each other team has their cones located in equivalent positions throughout the course.Figure SEQ Figure \* ARABIC 5: Cones on string1.1 ScoringPoints are awarded as follows:Airborne robots have a constant 3x multiplier when scoring conesFor each contiguous powered territory on the first level controlled at the end of the match, a team will be awarded ten (10) pointsFor each contiguous powered territory on the second level controlled at the end of the match, a team will be awarded thirty (30) pointsFor each contiguous powered territory on the third level controlled at the end of the match, a team will be awarded forty (40) pointsEvery ten (10) seconds that a team controls a first level territory will result in one (1) pointEvery ten (10) seconds that a team controls a second level territory will result in three (3) pointsEvery ten (10) seconds that a team controls a third level territory will result in five (5) pointsCompletion of an action for the first time will result in ten (10) points being awarded to the team responsible.An action consists of opening the doors that hold pins or changing the direction of the ramp. The direction of the ramp can be changed by moving an assigned colored soccer ball at the base of the second territory to its holding area around the corner of the second level. Clearly, airborne robots are given a strategic advantage, and additionally, autonomous robots receive a 5x multiplier.2.0 Competition RulesThe competition can be summarized with the following core rules:Each match will be seven (7) minutes longEach match will consist of four or fewer robots The team’s color cone is the topmost in the stack of cones for control of a territoryTeams cannot attempt to control a territory unless it would be contiguous The following rules are also created for team composition:A team may have a maximum of six (6) members, consisting of a single team captain and up to five (5) additional members.To be eligible to compete, each team member must be a registered student at a university during the semester during which the competition takes place (Spring), the preceding semester (Fall), or both.Only official team members will be allowed in the pit area during the competition – all others associated with the team will be subject to the same restrictions as regular visitors.Other persons besides the official team members may aide in the construction of a robot, but only the official team members will be recognized at the competition site.Optionally a team may have a faculty sponsor, corporate sponsor or a parent as an advisor in addition to the six official team members. A sponsor is meant to be a sponsor only. If there is suspicion that a team is abusing the sponsor privilege at the competition proper, the extra person will be asked to leave. The sponsor is not an extra team member. The sponsor is allowed in the pit area during the competition but may not assist in the repair, modification or construction of the robot.Airborne entries have special specifications. They can weigh no more than 15 lbs. and fit within a 3 x 3 x 3 cubic feet box. In general, quadcopters are much smaller than this, and weigh much less. Though building a larger quadcopter should increase stability, since maximizing horizontal geometry leads to greater stability. Any rotors must be made out of plastic and should not be able to cut through the safety netting which will be suspended over the course. The same battery rules apply to airborne entries. Since airborne entries could pose a greater safety risk, teams planning on submitting an airborne robot must contact the JSDC Rules Chair and Director by December 1st with a description of the expected weight, size, propulsion type, and any other specifications of their robot which will be helpful in determining what kind of risk such an entry might pose to the safety of spectators, competitors, JSDC officials and the game course. After the proposal has been received, the Rules Chair and Director will either approve or deny the entry by January 1st. In order for the proposal to be approved, it is necessary to consider some of the safety specifications:No explosives,acids,bases,flammables,chemicals,liquids,animals,projectiles,combustibles, lasers, buzz saws, drills, EMP’s, or unstable power supplies. In summary, nothing that will endanger the contestants, spectators, officials or the course will be allowed. The Jerry Sanders Design Competition Committee and Safety Officials have the final judgment on whether a device constitutes such a danger. Any inquires on the degree of danger a certain design presents should be directed to a member of the Jerry Sanders Design Committee.For public safety and to prevent damage to the course, each robot, whether piloted or autonomous, must have an easily accessible shut off switch on the robot or be able to stop on remote?command. It is also important to consider the types of cones that will be picked up. The cones are Adams brand cones, approximately 2 inches tall with a 2 inch diameter hole.Cones will be hanging from strings that span the corners of the course. The strings will be four (4) feet from the first level of the course and cones will hang one (1) foot down from the strings.The first time a team successfully removes a hanging cone from a string will result in ten (10) points being awarded to that team.Robots may not come in contact with other robots, or manipulate or seek to use cones other than their own corresponding color. Each time these actions are commited, personal fouls will be assessed. 5 personal fouls results in disqualification, and each personal foul carries -10 points. Additionally, pilots may not attempt to jam frequencies or interfere with other pilots during competition. These fouls can result in immediate disqualification.So, the advantages to using an airborne robot are not only the 3x point multiplier, but also the speed advantages and the ability to be able to avoid having to use directional ramps in order to enter higher level territories. In order to verify these advantages, the performance characteristics of the quadcopter will need to be experimentally verified.3.0 ExperimentIn order to later conduct a simulation and also analyze and compare potential strategies for winning the competition, it is necessary to conduct experimental trials to determine the performance attributes of a potential quadcopter that would be used in competition.Three different experiments were conducted. The first measured horizontal velocity over three separate distances, both taken from measurements on the above CAD diagram, that roughly corresponds to territory locations. The quadcopter was already hovering when the tests were conducted. The pilot would fly the quadcopter back and forth between the given distances of 20.5 ft, 30.7 ft, 47.7 ft. each time an observer, observed the quadcopter travel the given distance the time would be recorded. These times, over 19 trials for each distance, were modeled as distributions using Arena. The data for these horizontal velocity tests are as followsTable SEQ Table \* ARABIC 1: Data for Horizontal Distance Figure SEQ Figure \* ARABIC 6: Horizontal Distance Distribution 20.5 ftFigure SEQ Figure \* ARABIC 7: Horizontal Distance Distribution 30.7 ft.Figure SEQ Figure \* ARABIC 8: Horizontal Distance Distribution 47.7 ftThe second experiment measured vertical velocity ascending velocity over 5 different distances, 2.5 feet, 4.5 feet, and 3.5 feet. Due to time constraints in the facility where tests were being conducted, only 5 trials were conducted for each distance. However the means for each distribution were very close together. The experiment was conducted by marking the distances on a wall and mounting a laser to the quadcopter. The laser was aligned with the wall in order to measure the vertical altitude of the quadcopter. The data is as follows:Table SEQ Table \* ARABIC 2: Data for Vertical Velocity Figure SEQ Figure \* ARABIC 9: Vertical Velocity Distribution for fall from 2.5 ftFigure SEQ Figure \* ARABIC 10: Vertical Velocity for 4.5 ftFigure SEQ Figure \* ARABIC 11: Vertical Velocity Distribution 5.5 ft.The third experiment was testing landing accuracy, in order to determine how accurately the quadcopter could hover and approach a cone. A target was created and the quadcopter pilot attempted to land the quadcopter in the center of the target from distances of 2.5, 4.5, and 5.5 feet. Pictures were taken of each landing, and the distances from the center were recorded. 8 trials were recorded from each distance. The data for these trials is as follows:Table SEQ Table \* ARABIC 3: Data Collection for Landing AccuracyFigure 11: Landing AccFigure SEQ Figure \* ARABIC 12: Landing Accuracy Distribution 4.5 Ft.Figure SEQ Figure \* ARABIC 13: Landing Accuracy Distribution 5.5 Ft.4.0 Strategy Analysis It is necessary to analytically evaluate the feasibility of winning this competition with an airborne robot. In order to do this it is necessary to consider every possible combination of winning strategies, by developing a programmatic method for determining possible strategies. These possible strategies can be compared and evaluated via simulation of competition that will determine the best strategy for competition.In order to do this it is necessary to have an understanding of the expected aerodynamic behavior of the quadcopter. For brushless motors, the torque can be calculated as follows:τ = Kt(I-I0) , K = torque proportionality constant, I = input currentWhere K is the proportionality constant of the motor and I0 is the current when there is no load on the motor. The Voltage drawn by each motor can be calculated as follows:V = IRm+KvωOmega is the angular velocity of the propellers, Rm is the resistance of the motor, and K is the proportionality constant corresponding to back-EMF. We can calculate the Power consumed by each motor from P= IV, by solving the torque equation for current. If we assume that the motor resistance and I0 are negligible, and will not be large enough to affect flight time in a noticeable way, then we can calculate power as follows:P = (τKvω)/KtAdditionally, fromP?dt = F?dx , we can solve for P = Fv = Tv , where v is hover velocity.The hover velocity is equivalent to:vh= T2ρAIn this case, rho is the density of the surrounding air, and A is the area, r2, swept out by the propellers, the thrust, T, when hovering is equal to the weight of the quadcopter plus the weight of the payload carried. From this, it is clear that larger propellers coupled with thrust being held constant, equates to less power being drawn from the power supply, and thus higher levels of efficiency.Armed with some of the performance capabilities of a standard quadcopter, as well as the placement of all cones on the course, a strategy for winning the competition can be developed. First, given the fact that a team can only control a territory that is contiguous to theirs, the heuristic is that their contiguous path will not be able to intersect with another teams until the second level at least. Once a team controls, the third level territory then all second level territories are contiguous to them, and they can then attempt to play defense by placing cones on top of any territories in the second level that become controlled by the other team. Thus the optimal strategy is one that allows a team to reach the third level first, and control it as long as possible. In order to do this, one must consider the placement of the cones on the course, and pick up cones such that they will have the shortest mean time between cone placements. The total points at the end of a match can be calculated as follows:F = number of times 1st level territory is controlled for 10 sec.S = number of times 2nd level territory is controlled for 10 secT = number of times 3rd level territory is controlled for 10 secX = number of contiguous 1st level territories controlled at endY = number of contiguous 2nd level territories controlled at endZ = number of contiguous 3rd level territories controlled at endAlso add 30 pts for the first time the direction of the ramp is changed, including the 3x multiplierThen referring back to how points are scored, and incorporating the 3x multiplier, the total points at the end of the match is:Pts. = 3F + 9S + 15T + 30x + 90Y + 120Z + 30The constraints for this function are as follows:F ≥ 5S ≥ 1X + Y + Z ≤ 12We know this since we must control at least 5 territories before we can move into the second level and at least one in order to move into the third. We calculated an expected number of cones we can pick up to be 12. So the maximum number of contiguous territories controlled at the end is X+Y+Z=12. Figure SEQ Figure \* ARABIC 14: Decision Making TreeThe above diagram illustrates the decision making process that goes into executing a strategy. This illustrates how the goal is to first cover the 1st level adequately, in order to move into the second, and then into the third level.Figure SEQ Figure \* ARABIC 15: Operational ScenarioFigure 15 represents the operational scenario that was composed to represent the decision tree shown earlier. The interactions are between the user, the quadcopter and the remote controller. This scenario was created to design the experiments that will follow. The first part of the scenario involves the user commanding the quadcopter to turn on and ascend. The next steps are for the user to proceed to the doors and attempt to pick up a cone. In this particular case, the attempt is unsuccessful so the process of attempting to pick up a cone must be repeated until successfully picked up the cone. The user then commands the quadcopter to proceed to first level territory and drop the cone into a pin. This entire process of picking up cone and dropping onto pin is repeated until all contiguous pins are claimed in the first level. The entire process of picking up cone and dropping into pin is also repeated for the second level territory until all contiguous territories leading to the third level territory are claimed. By following the above decision tree, using the distances to each territory mapped out in the CAD diagram of the arena. We can use the mean times from our velocity distributions to determine an expected number of territories that can be controlled in a 7 minute period as well as the number of points that this would attain in a situation where there was no competition.Refer back to the equation for determining points scored:Pts. = 3F + 9S + 15T + 30x + 90Y + 120Z + 30Mean horizontal velocity of 3.91 ft. sec ( from velocity test)Mean ascending velocity is 1.14 ft. sec (from velocity test)Time to acquire a territory: t = 2d/3.91+ 2*4.5/1.14 +15This equation for time accounts for the distance there and back motion between territories as well as ascending and descending. This equation incorporates an assumption that picking up and dropping off a cone requires 15 seconds. This is optimistic; it may take as long as 45 seconds. Table SEQ Table \* ARABIC 4: Expected time to gain control of territoriesD is the distance in feet to territories in the arena. The time to gain control is estimated from the equation above. These distances represent the territories needed in order to control the third level. It takes 234.1 seconds to control the third level, out of a total of 420 seconds in the match. The remaining time can be used to play defense and prevent other players from entering the second level. The mean time to control a territory is 33.4 seconds, which leaves time for 4 more territories to be controlled in a 7 minute period. Thus, the expected number of cones is 12. In summary, because territories must be contiguous to other controlled territories, the strategy is to always approach the nearest territory. Once the third territory has been controlled then a defensive mode will be assumed, placing cones on top of other teams controlled territories if and when they enter the second level. Executing this strategy, without any competition taken into consideration will yield 1908 points. A mean time of 33.4 seconds to control a territory may be too optimistic. If it takes 30 seconds to pick up and drop a cone, then the mean time to control a territory is 48 seconds, which is enough to control 8 or 9 territories. This means the third level can be controlled, with two cones remaining to use defensively. In this case, the final point tally, not considering competition is 1,281.Figure SEQ Figure \* ARABIC 16: Graph of Points versus Time with No Competition5.0 Mission RequirementsThe Quadcopter shall compete in the Jerry Sanders Creative Design Competition.A strategic support tool shall be developed to determine and verify optimal strategies for winning the competition.The Quadcopter shall have a mechanism to pick up and carry cones, in order to place them on territories throughout the course.6.0 Functional Requirements The Quadcopter shall have the ability to transmit a live video feed to a laptop.The Quadcopter shall be able to remain in flight for at least 7 minutes at a time. A 3 minute buffer is desirable.The Quadcopter’s batteries shall be rechargeable between rounds.The batteries shall provide sufficient power such that, the flight time and consistent thrust to carry necessary payloads can be achieved.The Quadcopter shall lift the cone to a sufficient height, at least 2 feet above the ground.The Quadcopter shall place cones on a 1 inch diameter 8 inch long pin.The Quadcopter shall have the ability to collect cones semi-autonomously.The Quadcopter shall be able to fly beneath a 6 feet net at all times, and have the ability to fly as low as 2 feet for an extended period, totaling no more than 20 seconds.The Quadcopter shall have a carrying capacity of at least 1kg, for a camera, cone acquisition mechanism, and cone.The Quadcopter shall be transportable, able to meet requirements to be checked onto a plane, or in the trunk of a sedan.The Quadcopter’s Arduino system shall be compatible with stability (fly-by-wire) software modules.Functional Architecture7.0 Design 7.1 Design Requirements The Quadcopter shall fit within a 3’x3’x3’ cube.The Quadcopter shall weigh no more than 15lbs.The Quadcopter shall be equipped with a FPV camera.The Quadcopter shall have a mechanism for picking up Adams Saucer cones with a 2’’ diameter central hole and 2’’ height.The Quadcopter’s propellers shall be guarded such that they cannot damage the netting above the course.The Quadcopter shall be equipped with a mechanism for concentrating the thrust such that it will not blow cones away when attempting to pick them up.7.2 AlternativesFigure SEQ Figure \* ARABIC 17: Alternate QuadcopterThree different options for Quadcopter kits are in consideration, which will likely need to be heavily modified. The first is manufactured in China and sold by the DC Area Drone Users group. It is essentially identical to the 3D Robotics model, but has a cheaper frame and may not provide the stability that we are looking for. The second model is the A.R. Parrot Drone. This drone offers a very long flight time and is built on an open source android Operating System, however it cannot carry anywhere close to the payload that will be necessary for this competition. The 3D Robotics ArduCopter, though more expensive, offers a reliable shipping time and stable frame. It also can carry 60% more than the necessary payload, and is highly customizable.7.3 Weighted ValuesFigure SEQ Figure \* ARABIC 18: Trade off AnalysisThe cones are 2” Adams saucer cones, testing scenarios will be developed to compare each technology option quantitatively to determine their precision (accuracy of picking up cones). The methods will first be tested in place without being attached to a moving quadcopter. They will all be compared with the same amount of trials. If any of the methods are able to achieve at least an 80% level of accuracy then they will also be tested on the quadcopter in flight.The weights were chosen based upon the nature of the competition. Flight time and carrying capacity are highly important, and keeping weight of the robot down is also important to ensure a lengthy flight time. The abilities of the technology such as autopilot and camera are less important as they are only needed in this competition for lower level tasks, and most of the flight time will completed manually. Because all options are reasonably priced, cost is only of medium importance to us. The DC drone user groups, imitation ArduCopter is clearly the best choice.7.4 Parts ListTable SEQ Table \* ARABIC 5: Camera SpecsAlso being considered: using a Camera, for manually guiding the cone acquisition mechanism. The Go Pro is the most popular configuration, but it is heavy. Using a cheaper and lighter 3DR FPV transmission is preferable.KV850 Motor10 " Propeller11" PropellerPower Supply3300 mAh 30/45C4000 mAh 30/45C4500 mAh 30/45C3300 mAh 30/45C4000 mAh 30/45C4500 mAh 30/45CFlight time9.5min11.4min11.7min10.8min13min14.6minMixed Flight time12.8min15.4min16.2min14.1min17min19.1minHover Flight time31.1m37.7min44.2min30.2min36.7min41.3minBattery Weight558g678g762g558g678g762gEfficiency67.3%67.6%67.7%71.4%71.7%71.8%Payload Capacity1356g1380g1394g882g897g906gTable SEQ Table \* ARABIC 6 Alternative Camera KitsKV880 Motor10 " Propeller11" PropellerPower Supply3300 mAh 30/45C4000 mAh 30/45C4500 mAh 30/45C3300 mAh 30/45C4000 mAh 30/45C4500 mAh 30/45CFlight time6min7.2min8min7.7min9.2min10.3minMixed Flight time8.8min10.5min11.7min10.7min12.8min14.3minHover Flight time30.5min37min41.6min28.5min34.6min38.9minBattery Weight558g678g762g558g678g762gEfficiency67.3%76.7%67.7%79.3%79.7%80%Payload Capacity1549g1579g1596g1035g1054g1065gTable SEQ Table \* ARABIC 7: Brushless Motor AlternativesTwo different Motors (KV 880 and 850) are being considered with configurations of 10 inch and 11 inch propellers allow with 3 different power supply simulations. Ideally a simulation and test plan of all 12 of these configurations will be conducted to see which yields the best flight time for the desired payload. The issue is that, Quadcopters with higher thrust can carry higher payloads, but have shorter flight times. Larger power supplies can give a better flight time, but also add more weight. This creates a catch-22 situation.Part NameCostFunctionMotor 880Kv$96Motor Type: 3DR multicopters come standard with 850 Kv motors and 10" propellers, we upgraded to 880 Kv motors and 11" inch propellers for even more power.APM 2.6 Autopilot System$159.99It allows the user to turn any fixed, rotary wing or multi rotor vehicle (even cars and boats) into a fully autonomous vehicle; capable of performing programmed GPS missions with waypoints. Available with top or side connectors.3DR Power Module with XT60-Deans Battery Adapter$99.96The APM 2.5 Power Module is a simple way of providing your APM 2.5 with clean power from a LiPo battery as well as current consumption and battery voltage measurements, all through a 6-pos cable. The on-board switching regulator outputs 5.3V and a maximum of 2.25A from up to a 4S LiPo battery.11" Propellers$31APC Propeller pack. Packet includes one Pusher and one Normal plus adapter ring for each. APC 11x47 SFP Style.3DR ublox GPS with Magnetometer$89.99This new design incorporates the HMC5883L digital compass, providing a convenient method of mounting the compass away from sources of interference that may be present in the confines of the vehicle.Quad Copter Frame$99.99This new ArduCopter frame design incorporates several new features. The landing gear is moved further apart making the vehicle more stable during takeoff and landing. It provides a wide unobstructed field of view for mounting cameras under the vehicle.Replacement Kit$114.99Replacement Kit: added a replacement kit to get back in the air fast. The replacement kit includes two motors, two propellers, two arms, two legs, two electronic speed controllers, spare bolts, and extra spacers.Telemetry Set$85.99Telemetry Set: highly recommended adding a 3DR Radio Telemetry Kit to your autonomous vehicle. Telemetry radios allow your ground station computer to communicate with your aircraft wirelessly, providing unparalleled ease of use for viewing real-time flight data, changing missions on the fly, tuning, and using an Android tablet as a ground station. Select 915 MHz for the US and 433 MHz for Europe. Arrives pre-configured with RTF copter! The telemetry set includes an air module, ground module (USB connector), telemetry connector cable, two antennas, and a USB extension cable.20 Amp ESCs with SimonK firmware$103.96Features-Extreme low output resistance-Multiple protection features: Low-voltage, cut-off protection / over-heat protection-3 start modes: Normal, Soft, Super-soft -Throttle range can be programmed-Smooth, linear and precise throttle response.-Separate voltage regulator IC for onboard microprocessor-Supported motor speed (Max): 210000 RPM (2 poles), 70000 RPM (6 poles), 35000 RPM (12 poles)Radio Control$429.99Radio Control: Multicopters require a radio control system to be used as either primary or backup control. If you don't already have an RC transmitter and receiver, Spektrum DX8 will connect and configured to meet the requirements.FPV Camera + Transmitter$249.99OSD/FPV: Selected the 3DR OSD/FPV Kit to stream live video from your APM: Copter with superimposed live telemetry data. Great for an immersive flying experience with complete in-the-air data. This kit includes a Sony HAD 520 line camera, MinimOSD on-screen-display module, 5.8 GHz video transmitter and receiver pair, two 900 mAh LiPo batteries, and a specially designed Y-cable to allow you to use your Telemetry Radio and OSD using the same APM telemetry port. The 3DR OSD/FPV Kit comes pre-configured with RTF copter!Batteries$137.99Battery Kit: 3DR multicopters are powered by rechargeable lithium polymer (LiPo) batteries. Selected this option to receive a battery kit that includes one 3S 4200 mAh 11.1v LiPo battery, a LiPo balance charger, and a LiPo bag for safely charging and transporting the battery. We are receiving three batteries with our kit.R & D parts$300For testing the various mechanisms for 'cone acquisition' mentioned in requirements, buying cones identical to the ones used in competition, 3D printing parts, testing various propeller sizes and motor configurations.Total Cost:$1999.84Table SEQ Table \* ARABIC 8: Parts List Cost7.5 Pick up MechanismA mechanism is needed for picking up cones and transporting them within the arena. Currently there is no precedent, as previous airborne contest entries did not have much success with picking up cones, so it is necessary to test numerous options. The first option being considered is a remote 3-piece claw with high friction anti-slip surface that can be closed on the surface of the cone and can be guided into place with an FPV camera. It is possible that the quadcopter will not fly steady enough to manually pick up cones with a high level of precision, so the other option is to integrate open source autonomous color detection software with the microcontroller to automatically track the cone and grab it autonomously, once the pilot has manually positioned it within certain vicinity. The third option being considered is a “hook” that can be inserted into the central hole of the cone and then expanded. The issues with this are that it may not be reasonable to expect the quadcopter to be able to fly with the precision necessary to center the quadcopter; additionally it may not be possible to unstack cones that are stacked. The budget for testing and developing different configurations has been included in the R&D budget. Figure SEQ Figure \* ARABIC 19: Figure of Pick up Mechanism Possibilities8.0 Simulation DesignAs outlined in the strategy analysis section, since a quadcopter will likely need to ignore the cones that are behind the doors or hanging from strings, so the quadcopter must focus on the three stracks of cones that are available to them. One thing that is true for all competing teams is that they must only control territories that are contiguous with ones they already control. So heuristically, it can be determined that the only instances where this can occur, is when the contiguous shape created by one team approaches a territory that’s already controlled, or contiguous with a line of territories controlled by an opposing team. Since the most points come from controlling territories in the second and third level, the only place where this is likely to happen is inside of the second level. The team that will have the ability to play defensively is a team that is already controlling the single third level territory, since at that point all other second level territories are contiguous, and they can prevent other teams from controlling anything in the second level. Hence, the “name of the game” is to move into and control the third level as soon as people, hopefully first, and use the remaining time to place cones onto opponents controlled territories in the second level in order to prevent them from taking over the third level.Since the rules state that a team can only control a territory contiguous to a territory they already control, this means the first territory a team controls must be contiguous to their home territory in the first level. In this case, it was previously illustrated in the that at least 5 territories must be controlled in the first level, and one in the second before the third can be controlled. Choosing where to pick up cones on the course can greatly affect the time to control territories. For a quadcopter, the ideal strategy would be to immediately proceed to the nearest stack of cones after controlling a territory. For a ground robot that can open the doors, in instances where the nearest cones are the ones behind doors, picking those up may be beneficial.Now that experiments have been conducted, in order to model horizontal velocity, landing accuracy, and vertical take off velocity, these distributions will be used in the simulation trials, alongside the distances of cones and territories to be controlled in order to determine expected number of territories that can be controlled, and output the number of points. The landing accuracy distributions will factor in to determining error potential for picking up cones as well as estimating the time to pick up a cone. The strategic support tool will be designed using an object oriented language, with a robot class that takes parameters, for max velocity, agility, user error, and precision of picking up cones. During the simulation, user error and precision can be combined with a random number to determine whether or not a cone is picked up or not. These stats can be used to determine a maximum amount of cones a robot can pick up.Since a quadcopter will only be picking up cones that are in one of the three stacks, the time to pick up and drop off a cone will be the same. However for competing robots, it will take different lengths of time to pick up cones on strings and behind doors. Accurate estimations of this time will need to be factored into the simulation. Simulation trials will be ran against 4 competitors, and all of their possible choices for cone placement and choices of cones to pick up, in order to see the outcomes. In the simulation, whichever team is currently in control of the third level territory will assume a defensive role in order to prevent the other team from controlling the third level.Ideally this simulation will be conducted using MATLAB, and come complete with a GUI for user interaction and visualization of territories controlled by each team, as well as which cones they selected. Though if necessary, for web optimization, it could also be written in a more web friendly scripting language such as Python. The output will show the CAD diagram with highlighted controlled contiguous territories, as well as selection of cones.9.0 Project Management9.1 Schedule of Work The project schedule below shows the tasks and their expected durations. Figure SEQ Figure \* ARABIC 20Figure x shows the schedule of work and Gantt chart for the entire project. The Critical Path is marked in red. A critical path can be defined as a “sequence of stages determining the minimum time needed for an operation.” The tasks that are part of the critical path include: the context analysis, simulation design, mission requirements, modifications based on project briefing 1, physical construction of the quadcopter, testing and flying of the quadcopter, and all preparations of the final papers and presentations. Tasks that lie on the critical path need to be closely monitored to avoid any delays in the project.9.2 BudgetThere were a couple of budgets that were calculated for this project. These estimations are based on a salary of $55 per hour. An overhead price has to be adjusted since George Mason delegates 53 cents out of every dollar an employee charges their customer bringing the total to $117 per hour. The three budgets that were composed were composed by having a 10% overall budget expense and a 90%. The cost of the quadcopter itself was included as well as travel expenses to the Jerry Sanders Design Competition in Chicago. The blue line in the figure is the predicted total cost of the project, parts and labor. The hours worked was determined by predicting the number of hours each individual employee planned on working on the project. The red indicates the actual cost incurred up at this and will be updated every project briefing to show progress. At this point in time, no money has been spent on the parts; the majority of the cost has been labor. The total cost for all three budgets are as follows: $125,000 for planned cost, $137,000 for the worst case, and $113,000 for the best case. The following figures are breakdowns of the salary (table 9), conference costs (table 10), and all the calculations for the budgets (table 11).Table SEQ Table \* ARABIC 9: Hourly Wages Table SEQ Table \* ARABIC 10: Travel ExpensesTable SEQ Table \* ARABIC 11: All BudgetsFigure 10: All Predicted Budgets for 38 weeksFigure SEQ Figure \* ARABIC 21The Cost Performance Index (CPI) and Schedule Performance Index (SPI) for the project are calculated for each working week. A ratio closer to 1 is ideal and it shows whether or not the team is ahead in schedule or behind. It also states whether or not the team is under budget or over budget for the current week. An SPI greater than 1 is good meaning the project is ahead of schedule. A CPI less than 1 means the cost of completing the work is higher than planned, a CPI of 1 means the cost of completing the work is right on plan and greater than 1 means the cost of completing work is less than plan. A CPI greater than 1 is not necessarily a good thing as it shows that the initial plan was too conservative.9.3 Project and Risk MitigationRiskMitigationDamage during testing3D print parts for low level test configurationAdequate practiceDamage during competitionPurchase spare partsBudget for spare partsAdequate practice timeDamage during transportation of quadcopterTransport quadcopter using a reliable carrierDrive to competition and ensure delivery of quadcopterIncorrect simulation assumptions are madeEnsure assumptions used in the model are valid by analyzing prior competition resultsInvolve as many of the stakeholders as possible.Operator errorEnsure enough practice occursPractice with a small scale quadcopter Select the best operator Computer Error (if autonomous route is pursued)Ensure to the best of ability that no error is madeCreate operational scenarios to ensure that all scenarios are considered.Revert back to manual controlExtensive testingHardware Failure at any timePurchase extra parts Ensure adequate practice time Going Over BudgetPerform adequate analysisFollow CPI and SPI chartsTime ConstraintStart earlierWork efficientlyActively monitor project scheduleLoss of essential Quadcopter PartsKeep an active inventory of all active partsPurchase of additional partsTable SEQ Table \* ARABIC 12: Risk Analysis TableTable 8 is a table of risk analysis about potential risks that could arise at any point during the project and how to mitigate them.Damage during test: Having access to a 3D printer makes it easier to print out parts cheaply. Bringing the cost down and allowing more testing to be done. A lot of the parts on the quadcopter are interchangeable which make this a very feasible way to mitigate this risk. Adequate flight time and practice will ensure that the probability of a crash and possible damage during testing is reduced.Damage during competition: Purchasing additional parts will help in the case of damage during competition. It isn’t possible to use the parts that were 3D printed because it would influence the flight and other variables that shouldn’t be changed. Damage during transportation of quadcopter: Transportation of the quadcopter is something that was thought over a lot. There is a high probability of damage during transportation as nobody is sure whether or not carriers treat fragile items as such. Some ways to mitigate this risk is to either transport quadcopter using a reliable carrier which has some risk in itself, or a more feasible possibility would be to drive to the competition and that would ensure that our quadcopter makes it there in one piece. Incorrect assumptions are made: It is natural that there may be some incorrect assumptions made when thinking of an effective strategy. To mitigate, we ensured the assumptions that were used in the model were as correct as we could make them, another way was to involve as many stakeholders as possible to ensure that the assumptions that were made were correct.Operator Error: There is always a risk of operator error, whether it is in practice or during competition. Initial testing with a small scale quadcopter will help determine the best operator for the competition. Adequate practice time will ensure automaticity in flying the quadcopter minimize the risk of an accident during competition. Computer Error: It is very difficult to determine whether or not a computer error will happen if the autonomous route is chosen, but to reduce the risk the following things can be done: ensure to the best of the teams ability that no error was made when programming the microcontroller, create operation scenarios about a few possible scenarios where things can go wrong. Another feasible way to mitigate this risk is to revert back to manual control if any malfunction occurs. Extensive testing of onboard computer will minimize the risk of failure. Hardware failure at any time: Hardware failure can happen at any time during the project. To reduce this risk, it is necessary to purchase extra parts. With more test flights, hardware irregularities will become apparent.Going Over Budget and Time Constraint: All projects face a risk of going over budget and being behind in schedule. To mitigate these, it is important to perform adequate analysis to ensure all the parts that are being considered for purchase are essential. Another way is to follow the Cost Performance Index and Schedule Performance Index charts to ensure cost and schedule are kept at bay. To reduce time, it is important to start work earlier and efficiently as well as actively monitoring the project schedule at all times.9.4 Work Breakdown StructureFigure SEQ Figure \* ARABIC 22 Work Breakdown Structure10.0 ReferencesJSDC Rules List: , Dynamics, Simulation, Control - Andrew Gibiansky- for equations for simulation design ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download