Commercial vehicle drivers are governed by hours of service rules. These rules ensure drivers have sufficient rest to maintain safe driving habits. Many drivers choose to pull into a rest area to comply and Texas has areas specifically designed for truck parking. These areas become very popular at night and researchers have been investigating methods to keep track of available parking space and communicating the conditions to drivers before they arrive.
Traditional methods apply a count-in / count-out solution with the difference being the number of vehicles remaining in the parking area. This can give a count but there is no mechanism to self-correct for any errors made. Additionally, the errors are cumulative as time goes on until a ground truth count is made to reset the system. Another option is to install detectors at marked spaces but it is widely known that when conditions get busy, trucks park wherever they can.
Enter AI. Why not have a solution that counts the trucks within the whole facility rather than at an entry, exit, or marked space? Given proper views, an AI-based system can count the vehicles in the scene. The count can be used to generate messages for inbound message signs and even information directly sent to commercial vehicle onboard systems.
The following are a few examples of our research team’s exploratory work in this area.
