Useful Sensors and Boards for Robotics Projects
Disclaimer: This page may contain affiliate links. As an affiliate, I earn from qualifying purchases.
When creating a robotics project which you can use to gain experience and put on your resume, the first question is, what is your budget? I'll try to note the sensors worth getting at different budgets.
Free
Everyone loves free. I need to note this first, as it's worth seeing what kind of data exists before going out and buying new hardware.
My favorite autonomous vehicle dataset is NuScenes by Motional. They have production level sensor data from Boston, Pittsburgh, Las Vegas and Singapore, including 5x Lidar, 8x camera, IMU, and GPS along with their corresponding labels. If you're looking for a project to try out a SLAM algorithm or multitarget tracking algorithm, I would definitely recommend looking here first, as it will give you the best bang for the buck.
Cheapest
Here is where you start looking at projects involving some of the basic boards sometimes as early as Middle School or High School.
Boards:
Raspberry pi pico W: This board will need a laptop to connect, compile, and load code onto it. It is a very inexpensive board to start projects with. It can handle very basic code with just a few sensors, programming C++. The W means it has Wifi capabilities. I might use this project for home automation tasks, where one detects a door opening and closing, and another might keep track of the temperature outside.
Raspberry pi 3: This can hold a full operating system, generally you'll install Raspbian. It's a great way to start getting familiar with linux commands, connect a few basic sensors, and put together some basic programs. There are so many tutorials out there with project ideas, you'll be sure to find support along the way! Projects that a raspberry pi might be used for would be detecting your plant health or a miniature game console, or sometimes even creating a very basic wheeled robot. You can code in Python or C++. Be sure if you buy this to also get your MicroSD card and your power cable. You'll either want to ssh into your board or connect to a mouse and keyboard along with an HDMI cable to a monitor.
Arduino Uno: When I think of an Arduino I think of functional programming. If I sense this then do this. If you get this be sure to remember to also get the USB cable that connects from your board to your computer. You'll be coding on a laptop connected to the Arduino, compiling your C based code and loading it onto the Arduino, which will likely have your selection of sensors on board. You could create an automated dog feeder with this, a basic walking robot, an led matrix, or an amateur radio analysis kit.
Circuit Playground Express: This is an Adafruit board, and I have really enjoyed developing with it because there are so many embedded sensors, including motion sensor, temperature sensor, light sensor, sound sensor, it also has a speaker, 10 mini NeoPixels, a speaker, and push buttons. It is easy to attach alligator clips to add other sensors. This is very much a project and learning board, as I have had them stop working on me unexpectedly, but when they work they are great!
Sensors:
5MP Camera: Here is your basic camera you'll be able to attach to your Raspberry pi 3. The resolution is not great, but it'll be enough to start playing around with different Computer Vision algorithms.
Ultrasonic Sensor: This gives you a return based on the first thing the sound wave hits. It can detect up to 196 inches with an angle less than 15 degrees. This is often used to detect proximity, like when a robot is about to run into a wall, and gives a reasonably accurate distance away.
IR Infrared Obstacle Avoidance Sensor: This is also used for obstacle avoidance, but uses Infrared light for this detection. It has a range of 2-30 cm with a detection angle of 35 degrees. This sensor is less precise than an ultrasonic and is used more often simply to know if something is there.
PIR Motion Sensor Module: This is useful to detect human body motion using IR data. You might choose to use this for a home security system to prompt a light turning on, or a picture to be sent to you.
Mid Tier
Raspberry pi 4: This board gives a bit more compute than the Raspberry Pi 3. It will be more capable of running an object detection algorithm, though won't yet have enough compute to train an algorithm. You still have many GPIO pins where you can connect external sensors as well.
Intel RealSense Lidar Depth Camera L515: This is one of the cheapest Lidars you'll find. I used this for a Masters project where I detected the depth of objects in a room. A few things you should know are that this is a forward facing lidar, so you're only getting returns in a width by height format in front of the lidar. It is meant for indoor use only, as the sunlight interferes with the frequency outside. It was useful for my project, though it was difficult to work with as it's not very commonly used. The initial depth value returns didn't make sense to me, so what I had to do was look through their open source visualization tool, and leverage the converter function they use there to turn the initial return into an interpretable return of a certain depth. Honestly, I would likely recommend going with a different Intel Depth Camera over this one.
Intel RealSense D455: This type of camera is my favorite sensor to use for robotics. Intel has excellent support libraries. You have both camera and depth information, which makes your environment rich for algorithms, and if you have any movement in what it's connected to, you may want to get a depth camera with an IMU for better tracking algorithms.
Professional & Expensive
Nvidia Jetson AGX Xavier Development Kit: This kit has far more compute than a typical raspberry pi, and is often what is run in professional robotic systems. You'll generally run Ubuntu, and will have access to Nvidia GPUs for faster processing if needed.
Nvidia Jetson AGX Orin 64 GB Developer Kit: This is even more advanced and more powerful than the Xavier and is generally considered a top choice for a robotics development board. The Orin board would also generally run Ubuntu and will have access to Nvidia GPUs for faster processing if needed.
Ouster Lidar: This is a common industry professional lidar that is used on many production robots. They have different sizes, specs, and resolutions at different prices, but you should be able to find the right fit for your project here.