Navy

Navigation and obstacle avoidance for autonomous surface vehicles

Novel system uses an array of onboard sensors and sophisticated algorithms to identify and predict the movements of vehicles in the vicinity

Software & Information Technology

SSC San Diego Unmanned Surface Vehicle (Navy photo)

One area of hope for reduced accidents lies in vehicle avoidance systems now entering the market. These sensor driven onboard computers can detect other vehicles, lane markers, and road signs to alert the driver and even take control of the vehicle.

The Department of Defense is an early adopter of active avoidance systems and military research is funding extensive efforts to make improvements in vehicle safety. One such effort from the Navy is an object avoidance and autonomous navigation tool which incorporates a variety of sensors for both long and short fields of view. The core of the system – the computer processor – receives on-board sensor information along with map data to plan around potential collisions with more than 10 moving obstacles in less than 3 seconds. The internal sensors may include a magnetometer, LADAR, pan/tilt video camera, antenna automatic identification system (AIS), GPS, millimeter wave RADAR, RADAR, monocular vision camera, stereo vision camera, SONAR, gyroscope, compass, and accelerometer.

The system calculates a projected obstacle area (POA) – the area a moving obstacle may occupy at a future time period. The determination of a POA focuses on the time when a moving obstacle poses the greatest threat to a vehicle. The greatest threat to a vehicle is determined by initially finding the closest point of approach (CPA) – the shortest distance between two objects in time.

Components of the system include:

  • Deliberative obstacle avoidance – plans a path in the far field that follows the original path as much as possible using the Pathplannner 2D obstacle map
  • Reactive obstacle avoidance – responsible for avoiding obstacles that come in close proximity to the vehicle and primarily relies on the near-field sensors
  • Stationary obstacles – uses a map server to which users may add their own obstacles through operation or exclusion zones
  • Moving obstacles – the Pathplanner translates the third dimension of time of a moving obstacle to a 2D projected area for faster processing of vehicle location and collision point
  • Velocity obstacles – algorithm transforms a moving obstacle into a stationary one by considering the relative velocity and trajectory of the vehicle with respect to the obstacle and thus produces a collision area
  • Projected obstacle areas – predicts the area a moving obstacle could occupy in the future and calculates the closest point of approach which is the greatest threat to the vehicle

Do you have questions or need more information on a specific technology? Let's talk.

Contact Us