We present a robust path following control of skid-steered vehicles at
higher speeds on challenging terrain. First, a kinematic model with
experimentally identified parameters is constructed, in order to
describe the terrain-dependent motion of skid-steered vehicles. Using
the Lyapunov theory, a nonlinear control law is defined, guaranteeing
the convergence of the vehicle to the path. To allow smooth and accurate
motion at higher speeds, an additional linear velocity control scheme is
proposed, which takes actuator saturation, path following error, and
reachable curvatures into account. The combined solution is
experimentally evaluated and compared against two state-of-the-art
algorithms, by using two different robots on several different terrain
types, at different speeds. A Robotnik Summit XL robot is tested on
three different terrain types and two different paths at speeds up to
2.5 m/s. A Segway RMP 440 robot is tested on three
different terrain types and two different path types at speeds up to
6 m/s.
In the video, some representative experiments with the Segway RMP 440
robot are presented.