Arkansas Researchers Secure One Million Dollar Federal Grant to Automate Poultry Processing via Imitation Learning Robotics

University of Arkansas researchers develop ChicGrasp, an AI robotic system using imitation learning to solve labor shortages in the poultry processing industry.

By: AXL Media

Published: Mar 11, 2026, 5:53 AM EDT

Source: The information in this article was sourced from University of Arkansas System Division of Agriculture

Overcoming Labor Constraints Through Embodied AI

A research team at the University of Arkansas System Division of Agriculture has introduced an innovative robotics framework designed to stabilize the poultry supply chain against labor volatility. Developed in response to the workforce shortages highlighted during the COVID-19 pandemic, the ChicGrasp system represents a significant leap in embodied artificial intelligence. Unlike traditional automation, which often struggles with the unpredictable nature of biological products, this system interacts with its environment as an intelligent agent. Lead researcher Dongyi Wang, an assistant professor at the university, explained that the project applies the same imitation learning principles found in self-driving vehicle technology to the specific mechanical challenges of meat processing.

The Mechanics of the ChicGrasp Dual Jaw Gripper

The physical architecture of the system features a customized dual-jaw gripper equipped with precision pinchers designed to secure chicken carcasses by the legs. This hardware is tasked with lifting the birds and hanging them onto shackle conveyors, a repetitive and physically demanding role currently performed by human workers. Graduate student Amirreza Davar, who designed the gripper, noted that the system avoids the limitations of suction based or pre-programmed robotics. By utilizing a "diffusion policy" algorithm introduced in 2023, the robot treats control as a conditional denoising process, allowing it to adjust its grip based on the specific orientation and size of each individual bird.

Imitation Learning as a Path to Greater Accuracy

At the core of the ChicGrasp project is the concept of imitation learning, where the AI is trained by observing the trajectories of human teachers. This method bypasses the need for the robot to learn complex movements from scratch, providing it with a "ground truth" for how to navigate the processing line. According to Davar, this makes the system more efficient and accurate from its initial deployment. The AI stores camera inputs and movement data in a directory that informs the control of each joint in the robotic arm, allowing it to handle slippery, cold, and non-uniform carcasses that typically cause traditional robotic sensors to fail.