Introduction to Robotic Actuators

Introduction to Robotic Actuators
Robotic arm, courtesy of Jarmoluk via Pixabay

Robotic actuators are electromechanical devices that do some kind of physical task. They range in size from small consumer products to immense industrial equipment. Most robots contain multiple actuators.

A robotic actuator is a simple composition of smaller components.

Figure 2: System block diagram for generic robotic actuator

Okay... maybe not so simple. There's actually a lot involved in building an actuator. Required knowledge includes mechanics, electronics, programming, control systems, metrology, and more. Even for those who possesses the knowledge, every actuator is different. The myriad of different applications, operating systems, microcontrollers, motor technologies, sensor types, etc. make each robotic actuator challenging to design and integrate. Actuators are often designed by entire teams of engineers, each specialized in one core area of the actuator.


A choreographer composes actions and motions to achieve a high level objective, generally by orchestrating synchronous motion of multiple actuators.

The pervasiveness of video game controllers and high speed computing devices have made choreography equipment easy to obtain. The device you are reading this article on right now is likely a capable choreographer. Choreographers are often a combination of human input and software.

Live Choreographer

Human interface controllers allow direct control of a robot's motion through buttons, knobs, triggers, and other inputs. A human can choreograph as many actuators as their dexterity and cognitive abilities allow. RC vehicles (drones, planes cars, etc.) and Hollywood animatronics are often controlled through live choreography.

Figure 4: Youtuber James Bruton has built a live controller for animatronic eyes using joysticks.

Precomputed Choreographer

Precomputed choreographers allow for preparing coordinated motion of multiple axes in advance. They are often coupled with kinematic simulators to show a preview of the motion to be performed. Because the motion is precomputed, it can be iteratively tuned for very precise motion. CNC milling machines, laser cutters, 3D printers, and looped animatronics are often controlled through precomputed choreographers.

Dynamic Choreographer

Dynamic choreographers generate motion dynamically in response to sensor inputs or information from other robots. Modern dynamic choreographers are often coupled with advanced simulations (digital twins) to validate the robot's response in different environments.

Figure 6: NVIDIA has used digital simulations to make incredible advancements in robotic control technology.
Figure 7: A dynamically controlled robot that uses sensory inputs to accomplish a set of tasks to complete an overall objective. This robot was constructed by myself and fellow students during our undergraduate coursework.

Real-Time Controllers

Real-time controllers must maintain precise time keeping and synchronization to ensure that all actions commanded by the choreographer are performed at the correct time.

In early robots, Programmable Logic Controllers (PLCs) and electrical circuits were used to perform real time functions. PLCs are high-end controllers designed specifically for industrial automation; they cost hundreds to thousands of dollars. Electrical control circuits aren't terribly expensive but take a lot of specialized knowledge to design correctly.

Today, cheap, highly capable digital microcontrollers permeate the industry. An electronic chip smaller than a dime can replace what used to be an entire cabinet of electronics. They can be purchased off the shelf in low quantities for just a few dollars.

Drive Electronics

An electric motor typically requires a higher voltage than the real-time controller. The drive electronics use the controller's low voltage signal to control the motor voltage. The drive electronics are also designed to meet a motor's high current draw.

All motors must have their voltage switched within the motor's coils; this is called commutation. Brushed motors do this through mechanical means while brushless motors require advanced drive electronics to perform this task.

Electrical Sources

A motor and controller cannot be simply plugged straight into the wall. The wall voltage needs to be converted for the controller and motor by a power supply. The power supply needs to supply enough current for both. Common power supplies are intended for phones and computers... not robotic actuators. Power supplies come in all shapes and sizes, each with different features and drawbacks.

Electric Motors

Electric motors convert electrical current to mechanical torque. There are many kinds of motors, differing in their form factors, speed/torque capabilities, and price points.


Though electric motors generate torque, producing raw torque alone is often not enough for an actuator to perform a task; sensors allow monitoring and control of an actuator's position and velocity.

Switches are a basic but extremely common method of sensing an actuator's position. They can be used to detect when an actuator is at one end of its range of travel.

Figure 11: Limit switches are a very common end of travel sensor for an actuator.

Another common position sensor are encoders. Encoders sense an actuator's position without inhibiting motion. Encoders can also be used to estimate velocity by counting the position change over a period of time. Acceleration can also be estimated by counting the estimated velocity change over a period of time, however, the estimate quality degrades rapidly with each derivative.

Range finding sensors such as ultrasonic (sound) and lidar (light) measure distances from the sensor to other objects without touching them. They are not as accurate as contact methods like switches, but they can measure over a broad range of distance.

Figure 13: Ultrasonic sensors time the duration for emitted high frequency sound waves to bounce off objects and return to the sensor. They convert this time to a distance using knowledge of the speed of sound through air. Courtesy of Adafruit,

Accelerometers measure the acceleration imparted to the sensor, typically including gravity. This allows them to be used for global orientation measurement.

Though cameras are in nearly all modern consumer devices, using them as a sensor requires a fast controller and specialty knowledge. While most sensors produce just a few sensor readings at any given time, each pixel in a camera is a light sensor producing data. A $1$ megapixel grayscale camera is considered low quality by today's standards but still produces $1,000,000$ sensor readings at a time (one for each pixel).

Figure 15: Though cameras are accessible in laptops, tablets, and phones, integrating them into actuators requires specialized hardware meant for real-time sensing. Courtesy of Adafruit,

The possibility space for sensors is enormous. Other examples of measured quantities include voltage, current, temperature, pressure, humidity, strain, and infrared light. It would be impossible to cover them all here.

Mechanical Outputs

The mechanical output of a motor is the connection point between the motor and the rest of the actuator. The general slenderness of motor shafts together with the expense of shaft machining means that many motor shafts are poorly designed for transmitting the full torque capabilities of the motor.

Motion Transmission (Gears, Belts, Screws, Pneumatics, etc.)

Motion transmission components transmit torque from the motor to the part of the actuator performing the task.

Synchronous transmission components (timing belts, gears) ensure that the motor rotation is synchronized with the entire mechanism. Asynchronous transmission components (pneumatics, hydraulics) don't guarantee the same.

End Effectors

An actuator's end effector is responsible for performing the actual task. Material cutting, depositing plastic, lifting boxes, looking pretty... these are all purposes of end effectors.