John Wright (JPL Mentor) Pei Liew (JPL Summer Intern) Alex Sciuto (JPL Summer Intern)
The overall goal of this task is to explore new interaction methods for rover path planning for short and long distance driving. Large, multisensor datasets are available covering a large portion of Mars, some in the vicinity of the Curiosity Rover, some near the Opportunity rover, and some for other landing sites that have been considered.
Analysis and display of these datasets have been attempted with limited success and use of the data for path planning has proved difficult. In addition, planning a path uses a non-intuitive process of generating rover commands, simulating them, visualizing the results, and then tweaking the commands until the path looks correct. Also available is local terrain information in the rover's immediate vicinity. Work has been done to fuse the local and long-‐range data visually but little has been done to fuse multisensor datasets for path planning.
The primary research goal is to find an innovative, flexible, and powerful interaction paradigm for path planning. The path generation process must be influenced by traversability measures, incorporate and visualize multisensor data used for the traversability analysis, enable rapid path generation and comparison between alternatives, and export generated paths to mission operations tools.
The ultimate goal of this research is to develop a powerful path planning capability that can be incorporated into the mission operations system for future in-situ missions. Tools utilizing these techniques would support both short-‐range drive planning as well as long-‐range strategic traverse planning. It would enable visualization of multisensor datasets that affect the path planning process and utilize traverse cost measures to direct the plan and evaluate alternatives.
Coming soon... Please check back to my blog posts during Summer 2015.