Warehouse automation has become an essential aspect of modern logistics management. A critical component of warehouse automation is the path finding algorithm, which is responsible for calculating the route for automated guided vehicles to navigate the warehouse. However, conventional single-agent path planning suffers from deadlocks, resulting in a significant reduction in throughput. On the other hand, optimal and complete multi-agent path finding algorithms are slow and computationally expensive, making them unsuitable for warehouses with multiple uncertainties that require constant replanning. In this paper, we propose a novel scheduling technique to avoid deadlocks by introducing nonstop areas based on the dynamics of vehicle intersections. We define critical areas in the warehouse as nonstop areas and prohibit robots from stopping and waiting within these areas. The robot’s reservation length is dynamically extended to ensure that it reserves the entire path through the nonstop area, preventing it from blocking other robots moving through the intersection in other directions. The optimal positions of the nonstop areas were determined in two steps: the placement step to find where the deadlock occurred, and the trimming step to remove the excess nonstop area tiles. Simulation results showed 100% deadlock prevention and significantly improved throughput with the inclusion of the proposed nonstop area. Moreover, the addition of the trimming step further increases the throughput up to 12.27%.

This paper presents a precise localization technique for cooperative transportation robots and a work object, e.g., a logistic cart, that utilizes three-dimensional depth data and an optical image obtained by an external depth camera. Both the robots and the work object are simultaneously localized by the camera installed on the ceiling. The projective transformation and shape approximation are accurate enough for the robots to take hold of the work object without being disturbed by fluctuations inherent in the depth data. Evaluations show that the proposed method achieves a sufficiently small localization error of 11mm for the robots and approximately 30mm for the work objects on average. We also demonstrate that a prototype of the proposed cooperative transportation robot system can automate a transportation flow that consists of approaching the work object, holding it, and transporting it.

We propose a method to determine control input on the basis of minimizing the risk-sensitive cost function and show the results of an experiment in which the method was applied to a cooperative transportation robot system that we have developed. In the robot system, two robots hold a work object to transport without an external fixing device. The mechanism yields the force interaction between the robots and the object, which results in unexpected random errors in transportation. We add a stochastic term to the system model to describe such errors and solve the stochastic differential equation numerically to estimate the cost function. Utilizing the risk-sensitive cost function enables us to find the control input that balances efficiency and safety. The experimental results revealed that the chance of a robot colliding with an obstacle during transportation decreased from 90% to 7% compared with the optimal control.

Growing shortage of the labor force calls for automation of conveyance operations through the use of autonomous robots. These robots are difficult to implement in complicated facilities such as distribution centers and warehouses because of frequent changes in placements of conveyed goods and dollies. In this paper, we propose a conveyance system using multiple robots and external sensors. Our conveyance system offers adjustability to environmental changes and low cost by linking the robots and the sensors via a wireless network. Moreover, we introduce a prototype system of cooperative conveyance that controls multiple robots based on location information of the robots detected by a stereo camera.