Trajectory generation for vehicles moving with constraints on a complex terrain

Abstract

In this paper', the methodology of generating an optimal trajectory on a complex terrain for a specific vehicle is proposed. The possible paths are constrained by the Limitations on the terrain and the capability of the vehicle. To deal with these constraints, the notions of forbidden paint, forbidden direction, and forbidden path are introduced. After certain constants are specified, the method of dynamic programming is then invoked to find the optimal solution. If the target is beyond the maximal range of the vehicle, appropriate senice stations are selected by using the auction algorithm To speed up the computation process, the ideas of bi-spiral scheme and instant update are employed. With all the techniques at hand, numerical results show that the proposed method can generate the desired bajectory eiTiciently.

Distance indexed trajectory generation for a helicopter robot for programming by demonstration

“…Trajectory generation has been extensively investigated in the past to meet various objectives such as removal of spatial singularities [2,9,15], obstacle avoidance in dynamic environments [9], tracking accuracy and time optimization [8,16,13]. To remove spatial singularities, various interpolation techniques have been studied such as Bezier curves [9,17], B-splines [10,15,17], polynomial interpolation [17], arc-line trajectories [10].…”

“…Optimal control techniques not only calculate state trajectories ( ) t s but the control input trajectory ( ) t u also [13,18,12] for a given set of targets. Model predictive control (MPC) [6,12], dynamic inversion control [14], dynamic programming (DP) [13], mixed integer linear programming (MILP) [18] and sequential quadratic programming (SQP) [8] fall in this category. However, these techniques are computationally expensive and require an accurate dynamic model of the robot with all the details, which may not be available in many cases.…”

“…To ensure traversability of the reference trajectory, it is important to consider a dynamic model [8,[10][11][12][13] of the robot and various constraints such as safety and stability constraints and/or criteria of optimization [12,14]. It is impractical to consider all disturbances, such as wind and payload, which may affect tracking performance.…”