Recipient Organization
PURDUE UNIVERSITY
(N/A)
WEST LAFAYETTE,IN 47907
Performing Department
(N/A)
Non Technical Summary
Delicate fruits such as strawberries are, most commonly, harvested by humans requiring the availability of extensive seasonal farmworkers. The increasing labor shortage strains the U.S. fruit supply chains. Robotic technologies have been introduced in agriculture to help humans with fruit harvesting. However, existing fruit harvesting robots suffer from limited sensing capability - being incapable of sensing the quality and firmness of fruits. Furthermore, existing fruit harvesting robots mostly relied on bulky wheeled mobile platforms, being lack of the agility to travel across dynamic and uneven terrain in agricultural field environments. There is a critical need to develop an intelligent robotic system with rich sensing capability and safe and versatile locomotion for autonomous delicate fruit harvesting. In this proposal, we aim to develop an intelligent legged manipulation platform for harvesting strawberries to help farmers and farmworkers perform repetitive, tedious, and physically demanding delicate fruit harvesting tasks. Specifically, we will (i) design an adaptive robotic gripper with embedded vision sensors for multi-modal sensing, including for touch, vision, shape sensing, and proximity; (ii) develop perception algorithms for state estimation of strawberries, and design a tactile reactive controller for robust strawberry grasping; (iii) synthesize nonlinear locomotion controllers that ensure provable safety and versatility for our legged manipulation platform to traverse various uncertain terrains, and (iv) evaluate the proposed robotic harvesting system in both lab environment and commercial fields in California and Florida. The proposed research has the potential to alleviate the labor shortages in farming and agriculture. ?
Animal Health Component
30%
Research Effort Categories
Basic
40%
Applied
30%
Developmental
30%
Goals / Objectives
Thegoal of this proposal is to developa fully autonomous intelligent system with a multi-modal sensing enabled compliantrobotic hand housed on a safe and versatile legged mobile platform to release farmers/farmworkers from repetitive, tedious, and physically demanding strawberry harvesting tasks.We plan to attain the researchgoal by pursuing the followingspecific objectives:1.Design a Novel Adaptive Robotic Gripper with Multi-modal Sensing Capability.2. Investigate Multi-Modal Sensing and Tactile-Reactive Control for Robust Grasping.3.Synthesizing Robot Controller for Provably Safe and Highly Versatile Locomotionand Manipulation.?
Project Methods
1. We will develop an adaptive, compliant, and dexterous robotic gripper with multi-modal sensing capability. We will use kinematic synthesis and analysis asengineering guides to design the compliant gripper. We will integrate an embeddedvision sensor into each finger for multi-modal sensing.2. We will develop advanced perception algorithms to convert the embedded vision sensor signal to vision, proximity, touch, and proprioception, simultaneously.With the multi-modal sensing modality in the loop, we will develop alow-level controller for tactile reactive gripping control as well as a high-level motionplanner for diverse motion primitives of the robotic arm.3. We will draw upon nonlinear control theory to develop a robot control approach that simultaneously ensures the safety and versatilityof robot operation. This will be done by enforcing safety constraints (e.g., foravoiding adverse robot-plant collision) and explicitly handling terrain uncertainties(e.g., terrain deformation and unevenness).