GPU Implementation for Automatic Lane Tracking in Self-Driving Cars 2019-01-0680
Development of efficient algorithms has been the focus of automobile engineers since self-driving cars become popular. A major breakthrough was recorded when six autonomous cars successfully navigated an urban area to the finish line, out of eleven cars in the 2007 DARPA Urban Challenge, this achievement intensified research in this field with new startup companies in the field being birthed. Despite the good promises of self-driving cars, it is way behind being a perfect system because of the complexity of our environment. A self-driving car has to understand its environment before it makes decisions on how to navigate, and this might be hard because the changes in our environment is non-deterministic. With the development of computer vision, some key problems in intelligent driving have been active research areas. Leading vehicle detection and tracking is a central and critical task in intelligent driving system. The advances made in the artificial intelligence made it possible for car manufacturers to try it out on self-driving cars. No Artificial Intelligence system is perfect, there will always be some degrees of failure, but a failure in self-driving cars can be too costly. One of the areas we need to be prepared for is the potential of self-driving cars to drive off track because of bad weather or when there is no lane marking on the road. It has been also reported in the literature that lane tracking is a computationally intensive process. This paper is focused on implementing an efficient algorithm using Graphics Processing Unit (GPU) computing for automatic lane tracking.