Robotics Technology

Robots can be defined as intelligent machines that accomplish (usually repetitive) work that humans have done in the past. Except robots never get tired or bored, they can do work that are potentially dangerous, or that would kill for people.

Most robots use a large number of sensors and motors, especially as they try to replace human task-oriented actions. Multiple sensors in cars, phones, and the pervasive use of technology is changing the way we live in fundamental ways. Robotics has the potential to merge into every day technology to where it’s taken for granted, much like smartphones and television.

A robot can only perceive information when it has sensors; sense of touch is accomplished with feedback sensors, sense of balance is accomplished with gyroscopes and accelerometers, and so on. Movement is accomplished with multiple motors that can make extremely fine and precise repetitive movements that can consistently deliver the same action or series of actions with the help of encoders and other positions sensors. Pulling all of this together is the brain; which for robots is a central processor or multiple processors or controllers that are programmed by the robot’s creators.

Robots typically employ two types of motors: steppers and servos. Steppers are a specific type of brushless electric DC motor that can operate accurately without a position sensor for feedback; they can make precise, small movements. Servos are a type of DC motor that require a means to determine their present position in order to execute a precise motion. Servos are also better at continuous, high torque work at higher speeds than stepper motors.

Servo motors servos operate in a 180° field of motion; they complete 360° of motion by going forward 180° and then back 180°. Servos can be commanded to hold a position, even if physically pushed, and will continue to hold position against a force up to what their rated torque value will allow. Servo motors need a position sensor (closed loop control) to determine where the motor is with respect to the poles, and apply the right amount of current to achieve the commanded position. Servo motors are efficient because power to the motor is removed when the servo’s shaft is at the commanded position. Steppers do not remove power and operate at full torque.

A stepper motor will move a certain number of pole positions based on how far it can travel per pole position, for example. It does not have concrete proof that it has actually arrived at the expected position, unlike the servo motor. Nevertheless, stepper motors can be extremely precise.

Much thought should be given to how the application fits the motor’s purpose when choosing motors for robots. If torque must be controllable, then consider a servo motor. If a precise position must be held at a total standstill, a stepper motor may be the better choice.

Tactile sensing enables a robot to detect physical contact with the environment. These sensors can be piezo resistive, capacitive sensing, or elastoresistive tactile sensors, the latter of which are so sensitive they are used in minimally invasive surgery instruments that can literally feel inside a patient.

Sensors interface to a processor by a serial interface, analog interface, or binary input. Serial interfaces include SPI, I²C, and UART protocols. Analog-to-digital converters convert input from analog sensors into digital input for processing or for data collection and storage. For binary or single-switch sensors, general purpose IO (GPIO) pins can be used. Sensor input to a processor can be direct analog input (a real world sensor), digital (on/off input), or streaming data communicated to the processor via some kind of networked (e.g., Wi-Fi, Ethernet) or communications bus (e.g., serial or USB).

The microcontroller looks at the latest information supplied by the sensors, and perhaps also combining that with stored data or user input, makes decisions on what to do. These decisions are guided by programmed instructions stored in the microcontroller’s memory. Task-oriented robots such as those used in a manufacturing environment may perform tasks based strictly upon what it is programmed to do. Programming these kinds of robots is fairly straight forward; the robot literally carries out instructions as programmed. And any mistakes made by unforeseen situations, mechanical limitations, or made by the programmer (e.g., a misplaced semi-colon) are also carried out by the robot. Robots that are programmed to make decisions based on artificial intelligence or deep learning do not always use the same kind of straight-forward programming, and some make decisions based on extremely large data sets as they "learn" and thus self-program, by analyzing data that was previously identified by humans.