Research

Small Farm Harvesting Automation with Collaborative Robots

Deploying autonomous collaborative agents for lavender harvesting missions

We aim to develop a robotic system that can apply automation technology to farms without requiring structural changes, specifically targeting small and family farms. Our goal is to provide harvesting performance equivalent to human labor level through an autonomous robotic manipulator system. Autonomous navigation and coordination of heterogeneous mobile robot platforms for automated harvesting are addressed with traversability constraints, functional heterogeneity, and fuel constraints. This research is a joint project with the Automation in Manufacturing and Industrial Systems Lab at MTU.

Coordination for Multiple Tethered Underwater Robots

Entanglement-free navigation for tethered underwater robots

While tethered underwater robot has the most significant advantages in stable communication and power supply, considerable drawbacks exist to using cables, including drag forces on and caused by them. Our goal is to provide autonomous navigation for multiple tethered underwater robots without any entanglement of the tethers.

Task allocation

Assign the targets to the robots and find an optimal sequence of visit

Routing

find an optimal path with a given sequence of targets

Scheduling

determine arrival and departure times on the given routes to avoid collision and tether entanglement

Coordination for Multiple Heterogeneous Mobile Robots

Multiple Depot Heterogeneous Traveling Salesman Problems

We are interested in developing operational strategies for multiple heterogeneous mobile robots, such as having different average running speeds, payloads, sensors, etc. Our goal is to develop algorithms that find a route for each robot such that given missions are completed by at least one of the robots while the maximum travel cost is minimized. This objective leads to reducing the last job completion time, which is an important goal in many applications. We target the algorithms to be able to produce good approximate solutions within a reasonable computation time so that they are applicable to real-time operations.

So far, we have utilized primal-dual techniques to lighten the computational loads while mathematically approaching the problem, which can lead us to have better solution qualities. Starting from a multiple depot heterogeneous traveling salesman problem with the min-sum objective, we've worked on many variations of problems to deal with more general cases. You can find the most recent publication in the field here.

Three solutions were derived from different approaches for 3 robots with 30 tasks. The numbers next to the depots represent the index of the robots. The green, red, and blue paths represent the trajectories for robots 1, 2, and 3, respectively. The last task completion times for the min-max heuristic, min-sum heuristic, and LP rounding method are 8360, 11,445, and 12,309 seconds, respectively. The computation times for the min-max heuristic, min-sum heuristic, and LP rounding method are 2.43, 0.78, and 6495.11 seconds, respectively.

Reconfigurable and Scalable Microgrids Construction with Autonomous Agents

Deploying autonomous agents for target area mapping and energy generation

This research focuses on building a microgrid in an unstructured environment with unique mechanisms and controllers for power connector docking and an active cable deployment system. The robotic team consists of Unmanned Aerial Vehicles (UAVs) and Unmanned Ground Vehicles (UGVs) for mapping and power generation.

Page updated

Report abuse