Analysis of muscular activity of a cashier during different modalities of human robot collaboration

The effects of siding a collaborative robot (cobot) to a human operator on muscle activity during a scenario simulating a working task (i.e., a supermarket cashier moving cyclically packages) were investigated in this study, by considering different modalities of human-robot collaboration (HRC). Wearable sensors were used to collect kinematic and EMG data, and four different HRC interaction modalities were tested to determine their impact on the biomechanical risk associated with the examined working task. The range, amplitude, and smoothness values of accelerometer data coming from a sensor placed on the trunk, as well as coactivation indexes extracted from EMG data of different relevant muscles were calculated. Without cobot assistance, although a higher smoothness was found in the trunk movement, a high level of prolonged muscular activity was encountered usually associated to a potential increase in the biomechanical risk. In contrast, full assistance from the cobot resulted in the operator developing an intermittent motor control strategy, which reduces the muscular effort but losing smoothness in trunk movements. During two simulated tasks with a medium level of assistance, if the cobot imposed the cadence to complete the task, the required muscular effort remained constantly high, while if the operator was allowed to control the cadence, high smoothness combined with reduced muscular activity resulted. In conclusion, the use of cobots in repetitive tasks can influence the muscle strategy adopted by the operator during HRC. According to the analysis, the most effective approach is to reduce the workload of the operators while allowing them to maintain control over the speed at which tasks are completed.