Sensor-based Stooped Work Monitoring in Robot-aided Strawberry Harvesting

摘要

<abstract> <bold>Abstract.</bold> Manual harvesting of fresh market fruits and vegetables that involves sustained or frequent trunk flexion may lead to low back disorder (LBD), which is one of the top health-challenging occupational illnesses in agricultural workplaces. Conventional mechanical and robotic harvesters have not successfully replaced the judgment, dexterity and speed of experienced farmworkers at a competing cost. A long term goal of our work is to build agricultural robotic labor aids that achieve efficient, and ergonomically sound human-robot collaboration. This paper constitutes a first step towards this goal. A case study of robot-aided strawberry picking is presented, where small robots carry strawberry-filled trays from pickers to an unloading station, thus reducing non-productive walking time. When robots carry the strawberries, the possibility exists for pickers to pick almost continuously, thus maximizing their productivity. However, existing ergonomic studies have shown that the lumbar sagittal flexion becomes an increased risk factor for low back pain when its value exceeds <inline-graphic id=ID0ERFAE xlink:href=141913911_files/141913911-01.jpg/> for more than 5% of the working time. To prevent this condition, the inclinations of the trunk and the hip are measured in real time using wireless absolute orientation sensors and the lumbar flexion is calculated as the difference of these two angles and transmitted wirelessly to the robot. The robot can take suitable action to respond to this condition. In this paper, as a case study, the robot returns to the monitored worker to pick up his/her harvested crop only if the above condition has not been met; otherwise it stops a few feet away from him/her, thus forcing him/her to walk and ‘reset’ the prolonged stooped picking posture.