Speed Control of An Autonomous Electric Vehicle Using Fuzzy Logic With Computer Vision-Based Input

Abstract

A robust speed control mechanism ensures safety in an autonomous electric vehicle system. Such a system must dynamically adjust the vehicle's speed based on its surrounding environment. This research employs computer vision for object and road detection to measure the distance between the car and nearby objects. Fuzzy logic methods—specifically Mamdani and Sugeno—are utilized to automatically and stably regulate the speed of autonomous electric vehicles from their starting point to their destination. The control system considers various road conditions, including left-slanting, straight, and right-slanting roads, and the real-time presence or absence of objects. Testing is conducted across three real-world scenarios using distance and steering angle inputs. The servo angle represents the output, which ranges from 0 to 1800 and corresponds to the vehicle's speed. The results indicate that the Mamdani method provides greater speed control accuracy than the Sugeno method, which relies on a singleton output. For conditions involving left-slanting, straight, and right-slanting roads with objects within a 10-meter range, the Mamdani method produced outputs of 1370, 1800, and 1370, respectively, aligning well with predefined speed control rules. In contrast, the Sugeno method yielded 880, 1470, and 650 outputs for the same conditions, which did not adhere to the predefined rules for slow, medium, and fast speeds. In conclusion, the Mamdani method demonstrates superior accuracy and suitability for speed control in autonomous electric vehicles compared to the Sugeno method.