Optimal control technology applied to the motion planning of dynamic systems

Ref-Nr:

Technology abstract

 

A Spanish SME that is currently exploiting the Control Adjoining Cell Mapping with Reinforcement Learning technology. It allows to control any dynamic system, including complex real non-linear systems such as aircraft, satellites, RPAS and other vehicles. It is a technology that integrates systems dynamics techniques and intelligent learning schemes.
 

 

Direct contact: amari@kimglobal.com

- Andrea Marí -

Read more about this broker

Technology Description

The main application of CACM-RL is to serve as a technological baseline for seeking control algorithms with which to perfect the movement planning of the unmanned vehicle or system. Thanks to this technology, autopilots and controllers are not only able to generate trajectories, but also can optimize time, energy or distance.

 
CAPABILITIES

  • AUTOPILOTS FOR RPAS. Autopilots for RPAS designed by SOTICOL incorporate the latest technologies in sensing and instrumentation, as well as an innovative modular architecture of very small weight and high degree of integration, allowing it to grow in functionalities and operative capacity by adding advanced control modules.
  • OPTIMAL AIR TRAFFIC CONTROL AND OPTIMAL PLANNING OF TRAJECTORIES WITH TECHNOLOGY DETECT & AVOID. SOTICOL has developed a solid scientific foundation, integrating computational techniques of systems dynamics and intelligent learning schemes. This technique allows designing efficient algorithms of optimal control, able to realize an optimal management of air traffic, as well as allowing the aircraft themselves to generate optimal paths (4-D or 5-D) to their destination, making use of dynamic parameters. The generated trajectories must be characterized as a route carried out by the aerial platform between a "source state" and a "target state". That is, a state is a set of variables that can contain kinematic and dynamic information.
  • CONTROLLERS FOR LAND AND ROBOTS VEHICLES. The controllers for land vehicles and robots designed by SOTICOL have the ability to optimally plan their movement taking into account the kinematic and dynamic constraints of each platform, as well as include possible obstacles, either static or dynamic, that can be interposed in its movement. They are based on an innovative modular architecture with a high degree of integration, allowing simple growth in both functionality and new operational capabilities.
  • AUTO INDOOR & OUTDOOR. SOTICOL has developed an innovative indoor and outdoor auto localization system so that the platform that uses it does not have to rely on external location systems (eg. GPS). In this way, SOTICOL can integrate this capability into its autopilots and controllers to perform optimal control and planning, totally autonomous and without external dependencies.
  • SOTICOL GROUND CONTROL STATION. SOTICOL Ground Control Station (SGCS) incorporates a self-localization technology based on the sensor fusion between GPS positioning and IMU sensors, providing support to both aerial platforms (in approach and landing tasks) as well as to terrestrial platforms (in planning and control tasks). Aerial platform support is especially useful in those cases where the landing surface is moving (eg. the deck of a ship).

Innovations & Advantages

The offered products range from autopilots (SAPx family) with optimum control and planning for aircrafts, through ground control stations with encrypted & redundant communications, to solutions for optimum management of air traffic (swarming), Detect and Avoid (DAA), Auto-Location in absence of GPS, 3D discoveries of unknown environments, guided ammunition, gyro-stabilization platforms and Automatic Take-Off and Landing (ATOL) on dynamic surfaces.
 
The main advantage of the proposed technology is that it can self-adapt to changes in the controlled system. This means that, if there is a feasible solution, the controller can redefine its behaviour to take into account modifications in the system, the sensors or even the actuators. This characteristic makes CACM-RL a promising technique to build more robust and fault-tolerant systems. CACM-RL can be applied not only for building controllers, but also to generate planners with constraints and at the same time maximizing or minimizing energy, time or whatever goal considered. Finally, CACM-RL is not limited to linear systems it can be used in non-linear systems, providing optimal solutions for controllers and planners.
 
Additional advantages and innovations are the following:

  • THE MATHEMATICAL MODEL OF THE CONTROLLED SYSTEM IS NOT NECESSARY
  • OPTIMAL CONTROLLER 
  • VALID FOR linear, non-linear, stable and unstable systems
  • INDEPENDENCE OF THE SAMPLE PERIOD
  • OPTIMAL CONTROL IS ALWAYS APPLIED FOR EACH STATE 
  • REDUCED MAINTAINABILITY COSTS 
  • REDUCED INTERNAL PARAMETERS CAPACITY OF ADAPTABILITY TO THE PLATFORM

Current and Potential Domains of Application

  • Robotics
  • RPAS
  • UAV
  • Aircrafts