Path planning for robots by stochastic optimization methods (Q1264552)

The problem of adaptive trajectory planning for robots under stochastic uncertainty is considered; here, new information about the robots and their environment is presented on-line. Path planning for robots can be described mathematically by some variational problem. The authors use the methods of stochastic optimization for its solution. They consider two basic cases, namely, chance constrained problem and cost constrained problem. These problems are solved by means of finite element methods with B-splines, and by using neural networks. As a result, optimal control is calculated in real-time. Some numerical results are presented. They illustrate standard parameter optimization and NN (neural network) approximation.