Introduction to Self-Learning Adaptive Control Systems

Introduction to Self-Learning Adaptive Control Systems
Page content

What is Control System?

In general, a control system is a part of a larger system that manages the behavior of the larger system itself. The control system of a computer is the set of programs in the operating system that interacts with the CPU and peripherals in order to run the computer. In a CNC machine, the control system likely consists of a controller board (logic board) and a number of hydraulic actuators operated by control valve switches.

With the advancement of technology, control systems have become more sophisticated and intelligent. In the old days, mechanical switches were used for controlling a system, and then came relays, then programmable logic controllers (PLCs), microcontrollers, and now, microprocessors.

All the control systems can be broadly classified under two headers: sequential control systems and feedback control systems.

What is a Self-learning Control System?

The self learning or adaptive control system is a kind of advanced intelligent feedback based control system.

  • Actuators: This part of a system actually performs the output function of the system. For example, the pneumatic cylinder of the automatic door closing system is the output system. Hydraulic and pneumatic cylinders, electric and hydraulic motors are used as actuators in most systems.
  • System Performance Parameter Sensors: Performance of the actuators is sensed using these sensors to get the feedback signals about how much deviation exists in actuator performance. Mechanical sensors, laser-based sensors, and proximity sensors are a few names of the huge variety of the sensors used in industry.
  • Environment Parameter Sensors: These sensors are used for monitoring the change in the operating environments. The signals from these sensors help in refining the control parameters otherwise designed for the ideal operating environment.
  • Data Acquisition System: The signals sent by the System Performance Parameter Sensors as well as the Environment Parameter Sensors are collected and converted to useful data or knowledge by the data acquisition system.
  • Knowledge Base: Useful data or knowledge is stored systematically here. As the system matures, the size of the knowledge base increases and so does the efficiency of the self-learning control system.
  • Decision Making System: This is the brain. It sends optimized signals to the actuators based on the knowledge, experience, and present situation.

Applications of Adaptive Self-learning Control Systems

  • Active suspension systems: These intelligent automobile suspension systems use separate actuators to support the suspensions for the individual wheels. When a wheel rolls over a bump in the road, the control system senses it and makes a decision to release some pressure from the actuator connected to that wheel, which in turns allows the suspension of the wheel to rise, without disturbing the rest of car. When one wheel of the car finds a depression or pot hole in the road, the system increases the hydraulic pressure of that actuator in such a way that the actuator pushes the suspension of that wheel downward, and thus the rest of the car don’t get destabilized.
  • Auto-pilot or cruise control systems: The autopilot is an intelligent control system used mainly for aircraft to fly without the need for human interventions. The system continuously monitors the system parameters (like engine speed, vibrations, engine temperature, and airspeed) as well as the environment parameters (like altitude, humidity, and wind speed) and make optimum flying decision continuously. Similar systems are used for spacecraft and ships as well.
  • Adaptive mobile robots: The adaptive control system of intelligent mobile robots helps in acquiring and applying knowledge from the surrounding environment. It learns about things like new obstacles, and the motion of the robot keeps improving.


The self-learning adaptive control system is a kind of advanced intelligent feedback-based control system. It has complex feedback and environment sensors, a decision making system, a knowledge base, and work-producing actuators. The system makes intelligent operating decisions based on the current operating conditions and past experiences. This kind of control system is most suitable for applications where complete prior information of the time-varying control parameters is not possible. For example, the weight of an airplane continuously decreases as the fuel is depleted, thus increasing the range the airplane can fly on the remaining fuel. While designing this kind of system, it must be kept in mind that these applications cannot afford a dysfunctional control system, so system reliability is paramount.

Related Reading

Why are Control Systems Used?