Social robot navigation based on HRI non-verbal communication: a case study on avocado harvesting

To date, robotic applications in agriculture are still a challenging topic, which has been studied mainly for large farms. However, groves in particular, require tasks, such as picking and handling, that still require human labor force. In countries such as Chile and Peru, avocado is one of the main fruit production, but its growing in complex environments, making it difficult to fully automate the harvesting process. In this scenario, human-robot interaction (HRI) strategies can provide solutions to enhance the farming process. In this work, we propose the use of a HRI strategy via three visual non-verbal communication methods, with the aim of improving the avocado harvesting process leading to possible human workload decrement. Using such HRI directives, a robot motion controller is implemented for the robotic service unit to ensure that the interaction is socially acceptable during the avocado transportation task. The robot social navigation is tested in a simulated environment where the robot interacts with field workers to test three control tasks which are approaching, following and avoiding the human.