Probably the most widespread issues to do with a brand new robotic is allow autonomy, or at the very least get began on it with Simultaneous Localization and Mapping (SLAM). For years, the simplest method to do that in ROS has been to make the most of the open supply Gmapping bundle, a ROS wrapper for the Gmapping software program created by OpenSLAM. Gmapping offers a straightforward and comparatively simple strategy to start mapping and localizing your robotic.
Nonetheless, simply because one thing is free (or open supply), doesn’t at all times make it higher. With expertise utilizing gmapping ourselves, and listening to so a lot of our personal prospects and the higher ROS group lament the inaccuracy of gmapping for his or her initiatives, we knew there needs to be an alternate.
In response, final yr Clearpath launched the Autonomous Analysis Equipment (ARK) to the robotic analysis market. ARK is a software program and {hardware} bundle that makes use of the commercial grade autonomy software program developed by Clearpath’s materials dealing with division, OTTO Motors. The OTTO Motors workforce have put a whole lot of 1000’s of hours into creating, testing and optimizing, and the ARK is a analysis enabled, black-box model of that software program.
Whereas ARK has been obtainable on Clearpath robots for a number of months, our R&D workforce needed to pit it face to face with gmapping to find out simply how our proprietary localization algorithms stacked up towards the open supply incumbent.
Utilizing a Husky UGV with twin SICK LMS111 lidar, a number of exams had been run with each units of software program, together with guide and autonomous operation, after organising a VICON movement seize system as a floor fact.
By analyzing the ensuing datasets, we are in a position to see how every software program localizes the robotic in comparison with the movement seize knowledge. From this, we are able to decide which has the bottom error charges, in addition to see which one offers probably the most quantity of information factors.
For the total report with testing outcomes, in addition to entry to the collected ROS knowledge, click on the button under.
