DESIGN, SIMULATION AND TESTING OF SHRIMP ROVER USING RECURDYN

Transcription

1 Ready

2 12th Symposium on Advance Space Technologies in Robotics and Automation, ESA / ESTEC, Noordwijk, The Nethelands DESIGN, SIMULATION AND TESTING OF SHRIMP ROVER USING RECURDYN Shivesh Kumar, Raghavendra S, Mihir R Bhagat, Gangadharan K V, Department of Mechanical Engineering NATIONAL INSTITUTE OF TECHNOLOGY KARNATAKA Surathkal, INDIA

3 CONTENT Introduction - Shrimp - RecurDyn Work in RecurDyn - Modeling of Shrimp - Dynamic simulation on various terrain profiles Shrimp: Manufacturing Experimental Validation Results & Discussions Conclusions

4 INTRODUCTION Rovers: Most suitable for planetary exploration Over 300 design concepts developed already Shifting trends: Long-term, reusable rovers Rover deliverables increase: Long range Greater mobility Low power consumption High modularity Ease of control Fewer solo-missions. Regular inclusion of a greater bigger objective.

5 SHRIMP ROVER Innovative 6 wheeled rover Design at the Swiss Federal Institute of Technology (EPFL), Lausanne Design objective: Long-range mission for Martian exploration Image: EPFL Fully functional prototype was demonstrated at ESTEC (2000)

6 SHRIMP ROVER: KEY FEATURES 6 motorized wheels 2 parallel bogies, front and rear forks Passive control Excellent mobility: - Obstacle climbing - 2 x wheel dia. - Topple resistance - Upto 40 degrees - Pure steering possible (a) Image: EPFL

7 RECURDYN Multi-body dynamics software Advantages over other softwares: - No more excessive simplification - High solving efficiency - Good solving stability Equations of motion theory in recursive formation A product of High precision, fast solving.

8 MODELING OF SHRIMP Initial design based on SHRIMP III by BlueBotics ( Design made in CATIA V5 Imported to RecurDyn and joints defined Type of Joint Between body and wheels No. of Joints Type of Actuation 6 Active In front fork Passive, 1 Active Revolute joints In rear fork 1 Active In right parallel bogie In left parallel bogie 6 Passive 6 Passive Other parameters defined as well: Wheel speed: 30 rpm Friction: μstatic = 0.5 μdynamic = 0.3 Tire and road surface properties: Spring coeff. = 1000 N.mm Damping coeff. = 1 Damping friction coefficient = 0.9

9 DYNAMIC SIMULATION Kinematic and Dynamic simulation done using the multi-body dynamics solver of RecurDyn Simulation time and time steps chosen suitably The rover capabilities are tested on 3 types of terrains: - Ability to climb steps - Ability to climb a inclined surface - Ability to adapt passively with concave/convex terrains

10 CLIMBING STEPS Obstacle: Step Step size: 200 mm Under review: Torque requirements of the wheels

11 CLIMBING STEPS FRONT WHEEL REAR WHEEL

12 CLIMBING INCLINATION Obstacle: Slope Slope inclination: 40 degree Under review: Torque requirements of the wheels

13 CLIMBING INCLINATION FRONT WHEEL REAR WHEEL

14 CURVED SURFACE ADAPTABILITY Shrimp makes use of a passive suspension system Its is tested to check its adaptability to curved surfaces. Ex.: Concave

15 TORQUE ANALYSIS Wheel-Bogie Joint Position Max Torque (Kg-cm) Front Fork 32.1 Right Front 7.5 Right Rear 7.4 Left Front 7.0 Left Rear 6.8 Rear Fork 16.3 Maximum torque required is close to 32.1 kg-cm Factor of Safety: 1.25 Hence, Torque requirement becomes 40.1 kg-cm Available option: 45 kg-cm motors.

16 SHRIMP: MANUFACTURING Major Parts - Main body - Parallel bogies - Front & Rear fork - Electronics sub-system Built in-house at NITK Undergraduate project: Resources and facilities greatly constrained.

17 MAIN BODY Twin Alumnium pentagons form the base Supporting aluminium blocks: - To join the two plates - Support the load - Mounting point for revolute joints Bearings used for revolute joints Similar setup for the front fork

18 PARALLEL BOGIE Aluminium bogies mounted on either side of the rover Frames of C-section links that form a couple Mounted on freely rotating central pivot NOTE: Left and right bogie should be greatly identical to avoid mismatch in travel

19 FRONT & REAR FORKS Wheel mounted at the forks Front fork is similar to a 4 bar mechanism; wheel made to travel upward when it encounters obstacle Steering is achieved using servos whose axis is perpendicular to the ground and the axis of wheels

20 ELECTRONICS Key Features 2 parts: Base station & On-board control system Completely wireless control and data transmission using Xbee module Wireless video feed 2 ATMEGA16 processors in master slave configuration Light and portable LiPo battery along with battery protection circuit Interactive GUI using Matlab USB TV Tuner A/V feed RF Receiver Rover Control Commands User Sensor Data Live Video feed MATLAB based GUI Intel Core 2 Duo Processor (2.4 GHz) based PC FTDI based USB - UART bridge Wireless Xbee Module (Transmitter) Wireless Xbee Module (Receiver) Master Controller Atmega 16 based developer board (AVRiboard2.0) 3 units of: 8-28 V, 5A dual DC motor drivers with current sensing Right Bogie RF Transmitter Left Bogie Front & Rear Fork Front & Rear Fork s Servo for Steering BASE STATION Slave Controller AVRiboard 2.0 Servo Pod 2 servos SHARP IR Range Sensor Wireless Video Camera ON-BOARD CONTROL SYSTEM

21

22 EXPERIMENTAL VALIDATION Process Make the rover traverse obstacle paths to prove capabilities Measure armature current drawn by the motors and thus, estimate the driving torque requirements Obstacles Step Test: 8cm height Slope Test: 30 degree inclination Compare Driving Torque v/s Time graphs with those of RecurDyn

23 MEASURING THE TORQUE How to calculate torque? τ = Kt I

24 THE EXPERIMENTS

25 VALIDATION RESULTS Step Test Front Wheel Mean Absolute Percentage Error (MAPE) = 35.62% Slope Test Rear Wheel Mean Absolute Percentage Error (MAPE) = 42.87%

26 JUSTIFICATION FOR ERRORS MAPE typically varied between % for all cases. Effects of electrical DC drive model not included in our simulation. Non-linearities like BEMF voltage, friction between bearings, etc affect the system Incorrect data / estimation used in modeling: friction in revolute joints, contact friction, etc Mismatch between modeled and actual mass-inertia properties Simplified model: Nuts, bolts, clamps, etc not modeled

27 CONCLUSIONS Modeling & dynamic simulation of Shrimp through RecurDyn Virtual testing of rover on different terrains RecurDyn helped in selecting actuators for given payload requirements Experimental validation gave satisfactory results RecurDyn can be further exploited by better handling of problem at hand

28 ACKNOWLEDGEMENTS Mr. B Sridhar, Director of Function Dynamics India Pvt. Ltd. for issuing provisional licenses of RecurDyn for our project. NITK Alumni Association for providing a soft loan to attend this conference.

29 THANK YOU! Please contact: