Fuzzy based adaptive dimension parking algorithm including obstacle avoidance for autonomous vehicle parking NAITIK M. NAKRANI Electronics and Communication Dept. Uka Tarsadia University Bardoli, Gujarat INDIA MAULIN M. JOSHI Electronics and Communication Dept. Gujarat Technological University SCET, Surat, Gujarat INDIA Abstract: - Autonomous vehicle parking and obstacle avoidance navigation have drawn increased attention in recent times for autonomous vehicle-related solutions. Existing autonomous vehicle parking algorithms generally fail to mimic the human-like tendency to adapt naturally, and most of these designs are practically fixed. They do not preserve adaptive nature with machine dynamics, especially vehicles related. In this paper, a novel fuzzy-based adaptive dimension parking algorithm (FADPA) is proposed that integrates obstacle avoidance capabilities to a standalone parking controller that is made adaptive to vehicle dimensions in order to provide human-like intelligence for parking problems. This algorithm adopts fuzzy membership thresholds with respect to vehicle dimensions to enhance the vehicle's path during parking with taking care of obstacles. It is generalized for all segments of cars, and different simulation results are presented to show the effectiveness of the proposed algorithm. Key-Words: - Autonomous vehicles, fuzzy systems, hybrid intelligent systems, adaptive systems, dynamic path planning, ultrasonic sensing. Received: April 19, 2020. Revised: November 13, 2020. Accepted: December 1, 2020. Published: December 31, 2020. 1 Introduction Autonomous vehicle parking system [1]–[25] should be capable of parking a vehicle without any human intervention at the same time should be capable of parking in any environmental conditions. These systems should address designs and implementation related heterogeneous issues like trajectory planning, steering control, and continuous scanning of the environment for dynamicity. Literature suggests that researchers have worked on them individually or in a combination of a few of them as multi-tasking. The major challenge to build these systems remains the dynamicity of the environment and the feedback control. Behavior- based robot navigation architectures [11], [26]–[29] can be used as an integrated part vehicle parking system to track any unforeseen obstacle present in the surrounding. Navigation architectures required contactless sensing of the environment as an input, so many authors [10], [13], [23], [30]–[35] have used infrared sensors, ultrasonic sensors, LIDAR, laser sensors, and CCD sensor for their intelligent system design. Literature suggests that numerous algorithms for parking problems involved fuzzy logic theory because of its ability to use linguistic information required for complex systems to formulate a controller's rule base. Human-like intelligence can be easily mimicked by machines via a fuzzy set theory. For obstacle avoidance based navigation, linguistic information obtained from sensors is fuzzified to select membership functions and values heuristically, by experimentation or expert rules. However, most of the time, these fuzzy membership values are fixed for a specific set of models. An expert human driver can drive through optimal gaps with little sense of obstacles and understand vehicle dynamics. A drive-through experience is different whenever a person drives a WSEAS TRANSACTIONS on COMPUTERS DOI: 10.37394/23205.2020.19.33 Naitik M. Nakrani, Maulin M. Joshi E-ISSN: 2224-2872 277 Volume 19, 2020