Our lab develops tools and techniques for multi-robot collaboration aimed at automating construction site preparation tasks. Traditional methods rely heavily on human operators of heavy machinery. We focus on enabling autonomous earthmoving and material handling through aerial-ground robot coordination. Key aspects include:
- Accurate mapping and localization by UAVs
- Adaptive planning for UGVs based on updated terrain data
- Leveraging complementary strengths of air and ground vehicles
![Image credit: [Placeholder]](/assets/images/gen/projects/collective-robotic-site-preparation/bb_img11.jpg)
Aerial mapping
A UAV surveys a construction site from above to provide accurate mapping data
Projects
One project presented a computer vision method for UAVs to provide real-time localization assistance to UGVs engaged in site preparation. The aerial viewpoint enables tracking the ground robots and mapping materials. This data is integrated on the UGVs for more reliable autonomy. Another project introduced a toolkit called Shepherd for coordinating teams of UGVs through high-level motion planning and supervision. It allows non-expert operators to readily generate and simulate collective construction behaviors. The system was demonstrated through collaborative gravel manipulation experiments.
Outcomes
Our collaborative frameworks significantly improve the robustness and adaptability of autonomous outdoor construction robots. Aerial monitoring provides crucial mobility aid for ground vehicles. We continue to expand the multi-robot capabilities to handle more complex earthworks at larger scales. The ultimate aim is to reduce risk and maximize productivity for hazardous or remote site preparation tasks.