@phdthesis{Ampadu2025, author = {Ampadu, Kwame Owusu}, title = {Road traffic gesture autonomous integrity monitoring}, doi = {10.26127/BTUOpen-7222}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus4-72226}, school = {BTU Cottbus - Senftenberg}, year = {2025}, abstract = {Sensor enabled smart vehicles are predicted to succeed beyond human capabilities, thereby eliminating handling errors intrinsic to human driving. Driverless vehicles must be capable of detecting and responding to exaggerated ad hoc bodily movements that humans use to maintain road traffic safety. These nonverbal communication cues may be used to ease traffic flow, control infotainment systems or restore minimal risk conditions. The integration of sensor fusion techniques and high-quality benchmark datasets coupled to high throughput general purpose graphics processing units is spawning complex heavyweight algorithms useful for traffic scene understanding. These advanced methods are expensive in terms of costs related to hardware, software and energy consumption. Yet, the autonomous vehicle lacks the ability to participate in gestural exchanges. Following extensive evaluations of recent technological advancements in gesture detection and recognition in dynamic traffic environments, a new communication technology has been created for interaction between autonomous and traditional vehicles at unsignalized four-way intersections. This system operates without relying on most conventional navigation sensors and algorithms, allowing autonomous vehicles to communicate with traditional vehicles at intersections through in-vehicle traffic lights that utilize fuzzy logic. The algorithm implemented on low-cost microcontrollers can disperse road traffic and also perform extended blind spot detection. A single sensor is utilized, which remains largely unaffected by adverse weather conditions. The system requires no internet connectivity. It also avoids the need for large scale or long-term data storage, such as the driving cloud platform, ensuring backward compatibility with traditional vehicles. Simulations were conducted based on the order and direction of arrival at three junctions. Results have shown that autonomous vehicles at four-legged intersections, can now communicate with human drivers at a much lower cost achieving precise location classification and lane dispersion in under 30 seconds. This pioneering innovation presents an opportunity to turn every four-way intersection into a signalized junction at zero cost to governments.}, subject = {Advanced Driver Assistance System (ADAS); Autonomous vehicle; Fuzzy logic; Gesture recognition; In-vehicle traffic light; Fahrerassistenzsystem; Autonomes Fahrzeug; Fuzzy-Logik; Gestenerkennung; Fahrzeugintegrierte Ampel; Autonomes Fahrzeug; Fahrerassistenzsystem; Gestenerkennung; Fuzzy-Logik}, language = {en} }