Innovations in production logistics at BMW include the “Smart Transport Robot” (STR) developed by BMW in cooperation with the Fraunhofer Institute for Material Flow and Logistics IML. In the past, the STR navigated on fixed paths and required special sensors and a close-meshed network of anchors for localization. BMW wanted a solution that was more flexible, autonomous and economical. It was able to realize this with the IPA navigation software, which uses the safety laser scanner already integrated into the STR for navigation. Methods for dynamic path planning and obstacle avoidance assure the necessary degree of autonomy.
Implementing the STR navigation software was not without its challenges. This was due to the low quality and quantity of the sensor data paired with the fact that the robot has to navigate in environments which are sometimes highly-dynamic and highly-variable. In contrast to comparable robots, the STR solely detects its environment using a safety laser scanner with a limited field of view (180° view to the front). Furthermore, robot localization based on laser scanners is made all the more difficult because of the highly-variable environment with hardly any fixed structures, as well as the high dynamics due to forklifts, tugger trains and workers in the production halls. The Longterm-SLAM (Simultaneous Localization and Mapping) algorithm developed at IPA ensures reliable robot localization despite these difficult conditions. Based on current sensor information, the algorithm continuously updates the robot’s map of the environment and compensates for uncertainties and noise.
The software also meets BMW’s requirements as far as dynamic path planning is concerned. Here, the dimensions and safety fields of the robot must be taken particularly into account, and these vary depending on the load the STR is carrying and on its speed. Thanks to the software from IPA, the STR is still able to navigate through narrow gaps and avoid obstacles, such as vehicles which are parked or crossing in front of it. As a result, the STRs can act more autonomously, get stuck in impasses less frequently and reach their destinations faster. A further advantage of the solution is that it can be put into operation in new environments very quickly, a process which can be carried out by the BMW employees themselves. Using the Robot Operating System (ROS) in the software, application-specific adaptations or further developments, as well as interfaces to other components such as machine vision or fleet management, can be achieved without great effort.
Statement made by Josef Pilstl, Logistics Planning, BMW Group:
“Fraunhofer IPA not only develops software for prototypes, but also gives us reliable support during the STR's industrialization process, right up to the product's readiness for series production.”