Home
The main objective of this project was to develop and test a robotic system that could disassemble car door components autonomously, with the aim of supporting environmental sustainability and economic efficiency.
To achieve this goal the first step was to analyze a manual door disassembly process and gather information about key skills needed, to execute these steps. Based on this analysis, we developed a requirements specification, which is available here.
Based on the derived requirements, we structured the system into a set of functional modules that reflect the key aspects of the task: perception, motion execution, and coordination via a finite state machine (FSM). The perception pipeline processes sensor data to detect relevant objects (e.g., car door components, screws, tools) and estimate their poses, while additional steps provide context such as table height and subdoor positions. These outputs fed to the robot control system, which handles the execution of motion goals. The overall process is coordinated by an FSM that sequences actions such as initialization, object acquisition, picking, probing, and placement.
This pipeline is illustrated in the following diagram:
Find all the information on how we are controlling Eddie, the robot platform here.
The complete working pipeline is explained here