Kinematic models and motion control algorithms for a two-axle compliant frame mobile robot are examined. General kinematics describing the compliantly coupled nonholonomic kinematics are derived using velocity constraints that minimize traction forces and consider foreshortening of the frame. Given the complexity of these equations, the steering ratio a is defined to describe the relative heading angles of the front and rear axles. Simplified kinematic models are developed based upon a (Types I, II, and III) and the reference point used to guide the robot. Physical limitations and performance metrics (lateral mobility and maneuverability per unit of traction force) are derived to evaluate the models. Six groups of simulations and 24 experimental tests consisting of 120 trials evaluate the performance of the algorithms on carpet, sand, and sand with rocks. Results indicate that Type I (curvature-based steering) provides superior maneuverability and regulation accuracy, whereas Type II provides excellent lateral mobility at the cost of high traction forces, reduced accuracy, and potential singularities. Both models offer significant reductions in complexity for simplified control using standard curvature-based unicycle control algorithms. These results support expectations derived from performance metrics and physical limitations. Experimental results also demonstrate the efficacy of the robot to adapt to and maneuver over extremely rugged rocky terrain.

Posture stabilization of a compliant framed modular mobile robot is the subject of this paper. This is a new type of wheeled mobile robot that has advantages of a simple modular design that provides full suspension and steering capability without any additional components. Steering is achieved by coordinated control of the individual wheels to realize a desired trajectory. Due to the flexible nature of the robot, the kinematics is simplified by using an equivalent curvature based model, which increases mobility and decreases required traction forces for improved towing capacity. A time invariant control law is developed and extended to compensate for non-ideal initial conditions and drift. Simulation and experimental results are presented and show the proposed control law performs as expected.