Multi-protocol IOT Gateway

Intelligent Vehicles

Intelligent Vehicles

CarKyo -Machine learning planning - Project P-2

About CarKyo

  • CarKyo is a low-cost self-driving car solution.
  • Aimed at bringing self-driving technology to low-speed vehicles in a controlled area.
  • This means that our main challenges are to find low-cost solutions.
  • Typical applications of carKyo are in resorts like ElGouna

CarKyo - projects

  • For the year 2019, we have few support projects that are essential for the overall project and complement the current developments.
  • These are projects aimed at lowering the cost, enhance the overall solution.
  • The projects are chosen to provide a real contribution to the system.

CarKyo - technical areas of self-driving cars

There are 4 main areas:

  • Localization: Determining where the car exactly on the map.
  • Detection: Detecting obstacles and traffic signs
  • Planning: Getting the required destination and generating a proper path
  • Control: Execute the planning according to the car dynamics

CarKyo - planning project - P-2

Machine learning planning - Project P-2: 

  • Based on the target received from the central (fleet management) the car needs to find a global path (intermediate level planning) to reach the next stop provided by the fleet management, then determine a detailed (local path) that takes into consideration the car dynamics and the obstacles detected.
  • Currently, carKyo is using traditional AI techniques for planning.
  • Machine learning is a way by which the decision is based on experience. E.g.: providing how the human interacted with the problem before and imitate it.