Multimodal robotic last mile
MILE-v2 targets autonomous robots for smart last-mile delivery, which requires speed, efficiency, versatility, and payload capabilities. Future mobility hubs of AMAG and other partners require a solution for last-mile delivery that can carry freight, packages, tools, and material over long distances and challenging terrain. Our wheeled-legged robot is the only solution to carry payloads of up to 70 kg over long distances efficiently and fast while overcoming challenging obstacles. Thus, it outperforms state-of-the-art wheeled delivery platforms and lightweight delivery drones.
Based on extensive experience collected, we extend the robot’s autonomy skills for smart logistics as part of this grant and exploit our hardware and software development over the last five years. Supported by the mobility initiative, we aim to advance our solution into one of the most innovative autonomous last-mile logistics robots by including:
- Mapping and localization using multiple sensor modalities in indoor and outdoor environments
- Local planning with dynamic obstacle avoidance considering the robot’s unique mobility skills
- Autonomous shuttle services through:
- Teach and repeat
- Long-range navigation based on single navigation goals and pre-mapped digital twins
- Evaluation of overall logistics context
In this proposal, we focus on campus delivery tasks and proof of concepts of last-mile delivery applications like home delivery and freight transport. The developed technology can in the future become an integral part of Switzerland’s mobility hubs as a smart freight logistic system or find application in many other logistics applications. With a fleet of such robots, we expect to reduce traffic congestion, delivery durations, and CO2 consumption and provide automated delivery services that do not stop at the street. As part of the Swiss Autonomous Valley initiative, we created a digital twin of ETH’s Höoenggerberg campus, which will become the starting point of the targeted autonomy.
