I’ve had a few people reach out about what I would recommend for some further reading in robotics, and I’m compiling some of my recommendations here to share. The resources listed here are aimed significantly more at theory than practice; they are more focused on the algorithms to control robots and less designing and building robots themselves. These are some of the free educational resources that I’ve found most helpful over the years.
- Modern Robotics by Lynch and Park
This textbook, and the associated video lectures, are available for free courtesy of Northwestern University. This is a great introductory text that covers kinematics and dynamics for manipulators, path planning, contact mechanics, and more. Once you get through this text you’ll be prepared for more advanced topics in robotics. - Convex Optimization by Boyd and Vandenberghe
Robotics has steadily incorporated more and more concepts from the field of optimization over the last two decades. Though trajectory optimization is generally a nonlinear, non-convex optimization problem, most modern solvers operate by locally approximating the problem as convex and solving it iteratively. This textbook is an excellent introduction to the field of optimization, and will prepare you to understand trajectory optimization in robotics. - Planning Algorithms by LaValle
While all of the planning algorithms detailed in this text are worth learning about, this book contains perhaps the most important algorithm in the field of robotics: Rapidly-Exploring Random Trees. Probabilistic planning methods are an important tools for any robotics engineer to understand, and LaValle does an excellent job conveying both the theory and implementation of these algorithms. - Realtime Collision Detection by Ericson
One task that invariably shows up in almost every robotics context is collision avoidance. To avoid collision, it’s critical to be able to rapidly compute when two objects are intersecting. While this text is mainly aimed at video game engine developers, all of the lessons still apply to robotic simulation. Being able to rapidly compute proximity and intersection data between geometries is key to an efficient path planner. - Contact and Friction Simulation for Computer Graphics by Andrews, Erleben, and Ferguson
Another resource not explicitly aimed at robotics engineers, yet still incredibly informative. Contact and friction are necessary considerations when dealing with topics like grasping and manipulation, and they can be incredibly difficult to model accurately. Every popular robotics simulator has its own slightly different treatment of contact and friction (although some are more than slightly different, see Drake’s Hydroelastic contact model) but going through this course will give you a solid foundation to be able to understand any of them. - First Principles of Computer Vision by Nayar
Sensing is frequently incorporated into robotic systems, and as image sensors have decreased in cost they have become more and more prevalent. Computer vision algorithms are often remarkably simple, yet shockingly powerful. Nayar provides an excellent introduction to the camera model and image processing algorithms that is suitable to practitioners of any skill level.