We humans are naturally created robots with impressive capabilities. We can sense our environment, perceive and analyze it, and perform suitable corresponding actions. We continuously sense our environment using our five senses: sight, hearing, touch, taste, and sense of smell. These senses help us collect information from our surroundings and make sense of our environment.

Similarly, robots need sensors to gather information and stimuli from the environment to understand their surroundings. Just like our senses, the sensors of a robot convert the information in the environment that they are sensing into electrical signals which can be processed by the robot’s computer to perceive its environment and make informed decisions.

With hundreds of sensors available on the market, each designed to detect different environmental parameters, choosing the right one can be overwhelming. Today, we’ll focus on some of the most commonly used sensors in robotics. But before diving into the sensors themselves, let’s first understand the key parameters that define a sensor’s specifications.

Common Sensor Specifications

Our mobile phones and laptops have different specifications which we look at to identify the best product that suits our needs and fits into our budget. For example, if someone is looking for a gaming laptop then RAM, CPU and GPU specifications will be major deciding factors.

Similarly, sensors also have different specifications and understanding them and what they mean will help us identify the best sensor for our application. You can find the sensor specifications in its datasheet which is provided by the manufacturer of that sensor. Reading a datasheet might feel overwhelming since its flooded with numbers, tables and graphs, but let’s discuss a few important specifications to lookout for which will give us a good overview of our sensor.

Sensor Range

Sensor range tells us the maximum and minimum values of the parameter the sensor can measure. For example, a temperature might have sensor range 10-50°C which implies that the sensor can measure any temperature value which is between 10 to 50°C but it can’t measure any temperature value below 10°C or above 50°C.

Sensor Resolution

The resolution of a sensor refers to the smallest change in the input parameter that the sensor can detect and measure. For example, a pressure sensor may have a resolution of 0.5 psi, this tells us that if the pressure applied on the sensor changes by 0.5 psi or more then the sensor can detect that change and give the corresponding new pressure value as output, but if the difference in pressure applied is less than 0.5 psi, say 0.3 psi, then that change in pressure won’t be reflected in the sensor output.

Sensitivity

Sensitivity refers to how much a sensor’s output changes in response to a change in the quantity being measured, essentially the slope of the output-input characteristic curve. Output of a sensor with high sensitivity will change more drastically with change in input as compared to a similar sensor with low sensitivity.

Accuracy and Precision

Accuracy or absolute accuracy of a sensor refers to the difference between the value of the sensor reading and the actual value of the parameter the sensor is measuring. For example, a proximity sensor may have an accuracy of ±1% and let’s say that the distance that the sensor measured is 200cm, then the absolute accuracy implies that the actual distance will be between 198-202 cm (200 * 1/100 = 2, 200 ± 2 = 198 and 202).

Precision or repeatability of a sensor talks about how consistent the values of sensor readings would be if multiple measurements are made of the same parameter under similar conditions. For example, let’s say that our proximity sensor has the repeatability ±3 cm, this means that if we measure the same distance of 200cm multiple times, we will get sensor readings between 197-203 cm.

Electrical Specifications

• Operating Voltage:  This is the voltage range the sensor can operate normally without damage, for example, 10-30 VDC (DC Voltage) or 90-130 VAC (AC Voltage).
• Current Consumption – indicates the amount of electrical current the sensor draws from its power source, for example, a temperature sensor might have an operating current of 10 mA and a standby current of 1 µA. Maximum current specifies the amount of current that can be drawn into the sensor without damaging it.
• Power Requirements: The minimum power that the sensor needs to function properly, for example, 500mW. 

We have discussed some of the key sensor specifications here which might be helpful, there are many other specifications that a typical datasheet contains, but a full coverage of all the specifications is outside the scope of this article.

Sensors Commonly used in Robotics

Let’s look at some of the most commonly used sensors in robotics that are readily available in the market for you to get started with building your own robots! We will also touch upon the basic working principle of these sensors.

1) Proximity and Distance Sensors

A proximity sensor is able to detect the presence of nearby objects without any physical contact. Distance sensors can also measure the distance of the object in proximity from the robot. Ultrasonic sensor and IR sensor are the most widely used and cheaply available sensors for this purpose.

Ultrasonic Sensor

Ultrasonic sensors can measure distance of an object from the sensor by emitting sound waves which bounce off objects and the sensor detects these reflected waves. The distance between object and sensor is obtained by measuring the time difference between transmission of the sound wave and detection of the reflected sound wave and then multiplying the time difference with the speed of sound to obtain the total round-trip distance travelled by the sound wave which is then halved.

Commonly available ultrasonic sensor models: HC-SR04 and HC-100

Infrared (IR) Sensor

IR sensors work on the same principle as ultrasonic sensors, but instead of using sound waves, IR sensors use infrared light to detect objects. Light is emitted by an IR LED which reflects off objects and these reflected light rays are detected by a photodiode. The distance of the object from the sensor is estimated by measuring the time it takes for the reflected light to return.

IR sensors come in two types: Active and Passive. Active IR sensors both emit and receive infrared radiation, while passive IR sensors can only detect infrared radiation. For most applications, you will be using an active IR sensor.

2) Tactile Sensors

A tactile sensor detects and measures physical interaction, such as touch, pressure, and force, when an object comes into contact with it. These sensors mimic the sense of touch in humans. There are different types of tactile sensors available which work on different working principles.

• Resistive Tactile Sensor: It works by measuring the change in electrical resistance when pressure is applied on it.

• Capacitive Tactile Sensors: It works by measuring the change in capacitance when pressure is applied on the sensor.

• Piezoelectric Tactile Sensors: It uses piezoelectric materials which generate electric voltage when pressure is applied on it. The magnitude of this voltage is measured to calculate the pressure applied.

3) Inertial Measurement Unit (IMU)

An IMU measures the robot’s acceleration and angular velocity in 3 different axes (X, Y, and Z axes). This data can then be used to estimate the velocity, position and orientation of the robot in 3-D space. IMU’s are widely used in drones, mobile robots, autonomous vehicles and even inside our smartphones.

An IMU consists of a combination of devices, each measuring different environment parameters. A typical IMU consists of:

• Accelerometers: To measure linear acceleration. There are 3 different accelerometers to measure acceleration along the 3 different axes (X, Y and Z axes).

• Gyroscope: It is used to measure the angular velocity along the X, Y and Z axes.

• Magnetometers: (Optional) Measure the Earth’s magnetic field to determine absolute orientation (compass heading).

The acceleration data from the accelerometers is integrated to calculate the velocity which is further integrated to calculate position of the robot in the respective axes. The angular velocity values from the gyroscope are integrated to calculate the orientation of the robot in 3D space. Sensor fusion algorithms like the Kalman Filter are used to improve the accuracy of the IMU by combining the data from the different devices on IMU and other sensors to estimate accurate position and orientation values.

credit: https://www.tkk-air.co.jp/english/aerospace/inertial-sensors.html

4) Vision Sensors

Just like our eyes enable us to see, vision sensors are used to give the robot the ability to see the world around it. Vision sensors help robots perceive their environment by capturing images or depth information. They are crucial for tasks like object detection, navigation, SLAM, and manipulation.

Vision sensors are basically just cameras, but there are a number of different camera types available which suit different needs and purpose. Let’s look at the different types of cameras:

RGB Cameras (Monocular Cameras)

A monocular camera is a type of camera that uses a single lens to capture color images and videos. It is cheap and easy to use with low computational requirements. Most of the commonly used cameras have monocular vision.

Stereo Cameras

Just like we humans have two eyes, a stereo camera contains two cameras which capture the same scene from two different angles or viewpoints. This camera mimics the stereoscopic vision of human eyes which allows for precise depth perception, facilitating the creation of directly scaled 3D models of the environment with accurate spatial information. Stereo cameras are more expensive and need more computational power to combine and perceive the stereo image pairs using Epipolar geometry. They are especially useful for applications such as computer vision based autonomous navigation, visual odometry and 3D reconstruction of the environment.

Depth Cameras (RGB-D Cameras)

RGB-D cameras use cameras and other sensors to capture RGB images and depth information of the scene in real-time. Unlike regular cameras, which only provide 2D images, depth cameras measure the distance of objects from the sensor, creating a 3D perception of the environment. There are several different types of depth cameras available which use different technologies, but here we are going to discuss one of the best and most widely used depth camera which is the Intel® RealSense Depth Camera D435i. This camera uses the structured light approach for depth perception.

Credit: https://www.researchgate.net/figure/Location-of-the-experimental-field_fig1_359522270

The structured light approach is a depth-sensing technique that projects a known pattern of infrared light (such as grids or dot arrays) onto a scene and analyzes how it deforms upon hitting surfaces. A camera captures these distortions, and by comparing them to the original pattern, the system calculates the depth of each point using triangulation. This method enables the creation of high-precision 3D depth maps, making it useful in robotics, 3D scanning, facial recognition, and AR/VR applications. However, structured light struggles in bright outdoor environments due to sunlight interference.

These types of cameras are usually very costly and computationally expensive.

5) LIDAR

A LiDAR (Light Detection and Ranging) sensor measures distances by using laser light to scan and map the surrounding environment. It works by emitting rapid pulses of laser light towards a target and then measuring the time it takes for each pulse to bounce back. The sensor uses the speed of light to calculate the distance to the object, creating precise 3D models of the surroundings. LiDAR systems are known for their ability to provide highly accurate distance measurements, even in low-light or dark conditions, as they rely on laser pulses rather than ambient light.

The working principle of LiDAR involves three key components: the laser, the scanner, and the receiver. The laser emits millions of laser pulses per second in various directions. The scanner directs the laser pulses over a wide area, and the receiver detects the reflected light. By measuring the time delay between sending the pulse and receiving the return signal, the sensor calculates the distance to the object. This process is repeated rapidly to generate a detailed 3D point cloud, which represents the structure and features of the environment.

LiDAR has numerous applications in various fields, particularly in autonomous vehicles, robotics, and geospatial mapping. In autonomous vehicles, LiDAR is used for obstacle detection and navigation, providing a real-time 360-degree view of the vehicle’s surroundings. It helps the vehicle understand its environment, including the shape, size, and distance of objects, ensuring safe movement. In robotics, LiDAR is often used in SLAM (Simultaneous Localization and Mapping) systems, helping robots build and navigate 3D maps of indoor or outdoor environments. Additionally, LiDAR is used in surveying, topographic mapping, and archaeology to create high-precision 3D models of landscapes, buildings, and other structures.

Conclusion

In this article, we discussed some of the important sensor specifications and looked at five most widely used categories of sensors in robotics.

Getting started with robotics can feel overwhelming because there’s so much to learn. But if you take it step by step and focus on the fundamentals, you’ll build a strong foundation that will turn you into a skilled roboticist. Most importantly, have fun with it! Build as many robots as you can, experiment, and make them do all sorts of weird and creative things. The more you play around, the more you’ll learn.

Check out CodeRobo.AI to get started with your Robotics Journey!