Autonomous Logistics

Autonomous Logistics

Development of Autonomous Mobile Agent System in Complex Warehouse


Unlike vehicles with men aboard, autonomous vehicles require algorithms for path planning, static obstacle avoidance, dynamic obstacle avoidance, and many more. Especially, algorithms need to be considered for co-working with human workers to easily adapt the system in existing warehouses with lower setup cost.

 

Therefore TRUE lab is developing an optimized algorithm for each robot to find the shortest path from origin to destination for each mission while potential conflicts and deadlocks are prevented.


The goal of this project is to develop an algorithm which enables a large quantity of mobile robots (up to 300) operate in simulation. By having a testbed in a real warehouse and mobile robots specially produced for this project, in 2015-2016 the algorithm will be tested.

Development

Breadth First Search algorithm can be modified to be adopted in three-dimensional space with discrete time windows. Similar to Amazon's Kiva system, the algorithm generates paths for multi mobile robots, and check if any cell is overlapped by more than one robots. If there is a conflict, it finds another path which is the second shortest.This is repeated until all conflicts are resolved maintaining the shortest path as much as possible.

Simulation Example


Period: 2014.06~2017.05

Funding: KAIA

Team: Yuna Noh, Seyun Kim

R&D Partnership: Urban Robotics Lab (URL, KAIST), Korea Electronics Technology Institute (KETI), Nsquare, CJ Korea Express