Autonomous Bulldozer

This project involved the automation of a Caterpillar D11 tractor. The task of autonomous dozing is complex in large part due to the need to carefully control the load on the shovel to avoid stalling the forward progress of the tractor. For this project I implemented a memory-efficient personnel tracking system using active RFID tags that could run on the limited onboard computing of the bulldozer. Eventually we transitioned to more powerful computing which allowed us to embrace a computer vision approach that would eventually become part of the Caterpillar Detect product. This video shows one autonomous dozing pass of the D11.