Collaborative UAV Study Tan Han Rong, Ronald Department of Mechanical Engineering, National University of Singapore ABSTRACT This is a preliminary assessment into the feasibility of developing a UAV that can be used for collaborative teaming purposes. A tricopter configuration is investigated with the intention to develop a UAV suitable for operations in dangerous or hostile environments such as forests and urban areas. UAVs for such purposes are usually termed Micro Air Vehicles (MAV). The chief requirement is to keep the cost and complexity of the MAV low, in order to keep replacement cost low. Thus, the focus of this study was to assess the feasibility of using commercial off-the-shelf (COTS) components such as those used by Remote Control (RC) flight hobbyists. This MAV is also developed with the intention of it being a cost-effective platform for future undergraduate projects at NUS. LIST OF SYMBOLS M Moment MI Moment of Inertia α Angular acceleration T Thrust T o Thrust of un-tilted rotor m Mass of motor (= 0.045kg) g Gravitational acceleration (= 9.81ms -2 ) r Distance between motor and Center of Gravity (CG) r m Radius of motor (= 0.014m) M o Mass of tricopter D Drag force INTRODUCTION In recent years, there has been an increase in the need for intelligence gathering in challenging environments such as forests and urban areas. Such operating environments typically require a class of UAVs, termed Micro Air Vehicles (MAV) that are distinctly different in both size and configuration from conventional aircraft 1. This presents new technical challenges in areas such as control, low Reynolds number flight and lightweight sensors. In addition, the need for large numbers of cooperating MAVs and the high probability of losses during hostile operations dictate that MAVs for such purposes have to
be low-cost. An obvious solution would be to use COTS components but this would limit the level customization. This paper will study the feasibility of using off-the-shelf components in the control of a tricopter MAV. MAVs are becoming a commercially profitable area of research 2. A cost-effective MAV is necessary as a platform for future UAV research projects which are usually high risk. A key sticking point in many Final Year Projects (FYP) is that the cost of purchasing a conventional off-the-shelf platform is exorbitant, often taking a substantial part of the allocated budget. Such platforms typically require a skilled pilot whose services could take up another large part of the budget. An entire system purchased off-the-shelf simply does not allow enough flexibility for undergraduates to modify it in developing the various capability aims in a FYP. A MAV that is low-cost, simple to fly and easy to modify would be a boon to future FYPs. PLATFORM DESIGN A platform capable of hovering is required for intelligence gathering in confined environments such as forest and urban areas. A helicopter is a highly complex machine and a typical RC helicopter requires a very skilled pilot or a very expensive autopilot system. Low-cost RC COTS components have previously been shown to be incapable of controlling tailsitter MAV 3. A tricopter, based on a configuration experimented by a few RC hobbyists, is a suitable MAV that is simple to fly and modify, and will allow for the use of low-cost COTS components. The tricopter utilizes a control scheme that is based on the 120 CCPM mode of a typical RC radio transmitter. Where servos are typically connected to the receiver to provide pitch and roll control in a RC helicopter, electronic speed controllers (ESC) are now connected. The ESCs are in turn connected to the tricopters three electric motors. The pitch and roll for a tricopter are controlled by differentiating the rotating speed of the motors, thus varying the thrust of each rotor. A motor that can provide the required thrust was selected. The Turnigy Aerodrive C2826-1650 has been tested by other RC hobbyists to be capable of providing a maximum thrust of 0.5N with an eight inch propeller. An ESC is needed to control the rotating speed of the motor. The Turnigy 12A ESC was selected based on the abovementioned test results that the electrical power for the required thrust will not cause the current going to the ESC and motor to exceed 12A. The motor and the ESC are very much cheaper compared to brand name RC motors from the US and Europe, meeting the cost requirement. However, the thrust and electric power test would have to be verified. An additional risk factor is that the resolution of the ESC might not be fine enough to control the rotor thrust, and coupled with the short moment arm between the rotor and CG, might lead to a stability problem.
Table 1: Mass Estimation Component Mass (g) Motor (28-26-1650) x3 135 Battery 2200mAh 183 HXT 900 Servos x3 27 PG-03 Gyros x3 12 GY-240 Gyro 24 12A ESC x3 36 Structure 100 Total mass, M o 517 THEORY AND ANALYSIS Pitch and roll stability Pitch and roll in the tricopter is controlled by differentiating the thrust of the three rotors. In steady state, the moment arm of each rotor s thrust about the CG of the UAV should be equal. M = MI x α (1) Simplifying, T = mrα + mg (2) Differentiating, dt/dα = mr (3) From Equation (3), it can be seen that if the value of mr is small, a small change in thrust (dt) will cause a large change in α (dα). This implies that external disturbances to the attitude of the UAV will have greater effect, making it less stable. Thus, to increase stability, the mass at the rotor, m, or the distance between the rotor and CG, r, would have to be increased. The tradeoff for increased stability is that the UAV would be less responsive. Yaw stability Two of the tricopter s rotors are counter-rotating, thus only the reaction torque of one rotor has be countered. In the tricopter, all three rotors are tilted to counter this torque. An analysis is conducted to determine how much the rotors have to be tilted. Equating thrust and weight, 3Tcosθ = M o x g (4) Impulse-momentum principle, MI rotor x ω + 3Tsinθ.r = 0 (5) Simplifying, tanθ = [mr m 2 ω/2] / grm o (6)
From the technical specifications of the motor, the maximum rotational speed of the motor is ω = 1900 rads -1. Thus θ = 0.43. This tilt angle is exaggerated because it was assumed that the entire mass of the motor was rotating, when only the part of the outer casing is rotating in reality. This assumption was made as there is no way to disassemble the motor without damaging it. Also, the ω = 1900 rads -1 is the no-load rotational speed and in actual operation, the rotational speed is much lower. A tachometer would have to be used to determine the rotational speed of the motor in future work. However, this analysis does show that the tilt angle of the motors will be almost negligible in actual operation and there are no grounds for concern that tilting the rotor will lower the upward thrust excessively. Forward flight without pitching The original RC tricopter concept required the MAV to pitch forward in order to fly forward. While this was sufficient for hobbyists recreation, it will result in an operational MAV s sensors being pitched downwards where they will be rendered useless. Thus, it is proposed that two of the rotors be tilted to provide forward thrust. Equating thrust and weight, 2Tcosθ + T o = M o g (7) In horizontal plane, 2Tsinθ x cos30 = D (8) About CG, 2Tcosθ x r x cos60 = T o x r (9) Solving, T = ([M o g/3] 2 + D 2 /3) ½ (10) EXPERIMENTS This study is being continued as part of the UAV Forested Area Operations Final Year Project (FYP). As such, a thrust test conducted under the FYP verified that the motor and propeller will be able to provide the required thrust without exceeding the current limits of the ESC. A tethered flight test was also conducted, verifying that the COTS components are capable of controlling the attitude of the tricopter. Descriptions of experimental setup and test results will be included in the FYP report. CONCLUSIONS AND FUTURE WORK This study has shown that it is feasible to develop a MAV that is suitable for cooperative operations in challenging environments with low-cost COTS components. As part of a FYP, the tricopter will be developed further and adapted for operations in forested areas. The work done in this study has substantially reduced the proportion of the cost of the platform as part of the FYP budget. Areas for further development include the testing of forward flight without pitching, the integration of ducted fans and the inclusion of design features to enhance MAV collision survivability.
REFERENCES 1. Jane s International Defence Review, 05 June 2007, Straight from the forces' MOUT: UAVs enter urban environments 2. Jane s International Defence Review, 01 July 2006, Upping the stakes: demand rises for new-generation tactical UAVs 3. Tjin, Chia K.N. and Chan, FanTail Flight Controller Design and Flight Testing, RSAF Aerospace Technology Seminar 2007, Singapore 2007.