Evolutionary-based Coordination of Multi-Robot Systems with Dynamic Constraints; Effects of Automation in the Workplace: A Case Study of Amazon’s Fulfillment Centers

As robotic technology advances, automation is becoming increasingly popular in workplaces. Autonomous mobile robots are able to travel to specific locations in a complex system, solve a certain task, then move to the next assigned location. This allows for efficient production and resource allocation. Additionally, robots are able to travel and perform tasks that may be hard to get to or dangerous for a human employee to execute, thus increasing workplace safety. Many companies view robotics as an emerging and exciting opportunity to increase profits and optimize production, hence why more and more places are beginning to adopt this technology.
Purchasing the robots is one step of many for companies to begin implementation. My technical project, titled “Evolutionary-based Coordination of Multi-Robot Systems with Dynamic Constraints”, provides an algorithmic framework for how companies can optimally use and program robots. Along with two other systems engineering students, we examine the use of multiple robots in an industrial system to perform various tasks in an efficient manner. Our proposed framework begins with a robot mapping the desired environment using a scanner, such as Lidar. With the map, a user is able to provide several coordinates representing the tasks’ locations that the robots are to travel to and perform, along with the robots’ starting positions. Additionally, our algorithm accounts for user-input constraints to generate more realistic and optimal paths and assignments. These inputs are for priority, time and precedence constraints. The priority constraint allows the user to which tasks are more important and should be traveled to sooner. The time constraint accounts for how long specific tasks will take to complete. Finally, the precedence constraint ensures that certain tasks are completed before others.
The actual path generation and robot assignment was solved as an adaptive version of the Multiple Traveling Salesperson Problem using the Genetic Algorithm (GA). The A* algorithm is implemented to generate paths that minimize distances between the robots and each task. With the complete list of optimal paths, the GA is used to iterate through the possible path combinations and find the best paths for the robots to follow. The paths are sent to the robots controller which utilizes Artificial Potential Fields (APF) to move the robots. APF consists of three main forces: attraction to the goal, repulsion from obstacles and repulsion from other vehicles. With these three forces, the robots are able to travel to the goals while avoiding any obstructions and each other. This framework ensures that the robots follow optimal paths to their designated tasks, while avoiding obstacles and other robots along the way.
With the technical project focusing on implementing robots in an industrial system, it is important to understand the potential outcomes this may have on society. My STS research focuses on the effects of the increased usage of robotics in the workplace. More specifically, the effects of the use of robots in Amazon’s fulfillment center. This research examines how certain stakeholders, such as shareholders and corporate employees, fulfillment workers and customers, are impacted by this advancing technology. While robotics has led to many positive outcomes for Amazon, such as increased productivity, increased safety, and decreased packaging times, it has also led to several negative impacts. These consequences include employees having higher quotas for the amount of packages they need to handle per hour, employees fearing that robots could come for their jobs and shifts in labor skill demand. It is unlikely that Amazon will stop implementing robots, as they provide the company with many benefits, therefore, it is crucial that Amazon implements these robotics in a way that negatively impacts the least amount of people.

School of Engineering and Applied Science

Bachelor of Science in Systems Engineering

Technical Advisor: Nicola Bezzo

STS Advisor: Joshua Earle

Technical Team Members: Vihar Shah, Jose Vallarino, Patrick Sherman, Lauren Bramblett