Improving Operator Skills with Productivity Model Feedback

The performance of a mobile working machine is subject to operating conditions, operator’s actions, and technical condition of the machine. The ability of the operator has proven to be a significant factor when considering productivity or fuel efficiency. If the machine is in good technical condition with controller parameters tuned properly, the only way to increase performance, that is, productivity and fuel efficiency, is to improve the operator’s skills. The goal of this paper is to research the operator evaluation problem in the case of forest harvesters. The productivity variations of the machine between work shifts are modeled using variables that describe operating conditions and the performance of the operator in different work tasks. An adaptive-network-based fuzzy inference system (ANFIS) is proposed to model the productivity. The model is trained and validated using data from several operators measured in normal work environment during several months. An algorithm based on the gradient descent rule is proposed to give feedback about the most significant areas of improvement potential. The use of the gradient-based technique in offline analysis of the operator’s performance and work style is described. The variation of the performance between the operators is analyzed and the results are discussed.