Research M2 Open IPC
Intelligent Sensing and Safe Motion Planning
Scene awareness and safe motion planning are treated as inseparable parts of adaptive industrial robotics.
Sensor-fusion concept
Radar provides frequent scene updates, while 3D cameras provide geometry and semantic information. Fusion supports tracking, registration, segmentation and object-pose updates.
Safe planning
Proactive planning prepares safe alternatives before conflict appears. Reactive planning repairs motion when the scene changes. Collision-free motion alone is insufficient because industrial tasks also need safety-aware and human-aware costs.
How it works
Planning stack
- Offline and asynchronous Global planning and digital-twin work explore alternatives before runtime pressure becomes critical.
- Runtime and synchronous Path repair, velocity adaptation and micro-interpolation respond while execution continues under safety constraints.
Planning techniques
Informed sampling
- Global exploration.
- Local improvement.
- Adaptive exploration and exploitation.
- Multiple alternative paths.
MPC
MPC contributes trajectory generation under joint, velocity, acceleration and workspace constraints. It is expected to provide smooth feasible references while reducing jerk or motor load where possible.
ActivityA3 · A4
Activity A3 focuses on sensing for context awareness, including registration, tracking, segmentation and object-pose updates. Activity A4 focuses on local and global autonomous motion planning, including safe planning, informed sampling, runtime adaptation and MPC micro-interpolation.
Proposed targets
Project target 01
Autonomous joint and Cartesian planning.
Project target 02
Reduced interaction execution time.
Project target 03
Improved scene registration and tracking.
Project target 04
Safe velocity adaptation.
Targets describe planned outcomes, not results already achieved.