Sohan Saha Poddar (Matrikel-Nr. 403105)
A Bio-inspired Motivational Decision Making System for Social Robots Based on the Perception of the User
Biologically motivated models aim to solve various problems by taking inspiration from nature. It aims to create machines that interact with its environment by mimicking living organisms including humans. Biologically motivated models from the perspective of cognitive psychology take the process one step further by allowing machines to become intelligent, think and behave like humans or animals by incorporating underlying cognitive processes required for specific behaviour. It finds application in the area of cognitive robotics. The aim is to create robots which can assist humans in their lives, therefore, it’s desirable to create robots which are similar to humans or animals in behaviour. So studies are conducted by programming robots to simulate various cognitive processes to behave like a living organism. This can eventually lead to a better understanding of the cognitive processes and more usable robots.
In recent times there has been a trend to create robots which behave like living organisms. This journal presents an autonomous decision-making system based on a cognitive factor “motivation” and the perception of the human user used to train the system. The system is based on animal behaviour to do something based on the requirement to satisfy a drive or need. The main idea is that any living being’s brain strives to maintain an internal equilibrium of various physiological processes. In case there’s an imbalance the brain produces a drive which causes the organism to take actions which restore the balance. The external stimuli and the strength of the drive together determine the level of motivation to perform a task. In case there are several drives competing against each other the strongest one prevails. The robot was given two drives i.e social and relax. The social drive motivates the robot to interact with the user and the relax drive motivates it to sleep. The robot is equipped with a 3D camera to detect the user and also given the ability to perform several tasks like playing games with the user, sleeping, seeking attention from the user etc. However, it is not explicitly instructed which tasks would satisfy a particular drive. It gathers that knowledge in the learning phase. The robot’s training is user dependent. During the evaluation phase, the robot’s main goal is to maintain it’s well-being by keeping its drives to zero. In the actual experiment, the robot performs satisfactorily by maintaining a low and steady value of its predefined drives. The idea is very novel and opens the gateway to create robots which can behave like humans. It also raises a very important question of how to control robots which mimic human behaviour because such robots may learn negative aspects of human behaviour.
Journel: Maroto-Gómez, M. et al., (2018). A Bio-inspired Motivational Decision Making System for Social Robots Based on the Perception of the User. Sensors, 18(8), 2691.