Efficient Implementation of Parallel Path Planning Algorithms on GPUs

In robot systems several computationally intensivetasks can be found, with path planning being one of them.Especially in dynamically changing environments, it is difficult tomeet real-time constraints with a serial processing approach. Forthose systems using standard computers, a promising option is toemploy a GPGPU as a coprocessor in order to offload those taskswhich can be efficiently parallelized. We implemented selectedparallel path planning algorithms on NVIDIA's CUDA platformand were able to accelerate all of these algorithms efficientlycompared to a multi-core implementation. We present the resultsand more detailed information about the implementation of thesealgorithms