Real-Time Stochastic Optimization for Energy-Efficient Trajectories
Published in Proc. Robotics: Science and Systems Conference Workshop on Robot-Environment Interaction for Perception and Manipulation (RSS), 2016
Recommended citation: D. Jones and G. Hollinger, "Real-time stochastic optimization for energy-efficient trajectories," in Proc. Robotics: Science and Systems Conference Workshop on Robot-Environment Interaction for Perception and Manipulation (RSS), Ann Arbor, MI, June 2016. http://research.engr.oregonstate.edu/rdml/sites/research.engr.oregonstate.edu.rdml/files/real-time-stochastic_-planning-final.pdf
We present an iterative optimization algorithm for path planning. The algorithm samples smoothly deformed paths around a current best path and then updates the best path guess based upon a given cost function. We apply this algorithm to the problem of finding an energy-efficient path in an underwater environment. Results are shown for both a simulated current environment and using a Regional Ocean Modeling System (ROMS) ocean current data set. These results show that our algorithm is able to plan more feasible energy-efficient paths than current methods.