The self-balancing robot project is a fine way to start understanding the principles and methodologies that go behind the idea of maintaining balance. Here, I showed a simple robot chassis design, incorporated with the Arduino and the MPU6050, followed by a control mechanism using the PID to demonstrate that instilling balance in any robot is a highly specific task. And the level of difficulty only gets harder with the increased complexity of the robot’s design. Besides balancing, the new age robots are aimed to perform relevant tasks, whose success is determined by how well a robot can balance itself. 

Having completed the two-wheeled robot, we can now explore further by venturing into other categories of grounded mobile robots; the wheeled and the legged robots. Any multi-wheeled robot (with more than two wheels) appears to be perfectly balanced, unlike the two-wheeled one, which was susceptible to the challenge of tilting easily. However, it is important to remember that wheeled robots have to traverse through varying terrains and need to adjust their positions accordingly to not fall off-balance. 

One prospective project that we could attempt to work on is the case of determining the balance-restoring mechanism in multiple wheeled robots. In this case, the robot is allowed to fall and then has to figure out the position that it has fallen in. Based on that, an actuation process needs to be initiated to push the robot off the ground until a certain distance till it can return to equilibrium. The following questions can be asked concerning this project:

  • Which method of actuation would be the best? Pneumatic, hydraulic or electric?
  • How much force is required to push the robot off the ground?
  • Will the material used for actuation sustain the weight of the robot as it is being restored to its stable position?
  • Can the method of actuation be made universal for any type of terrain? For example, hard ground versus sandy area.

Imagine how disastrous it would be if NASA’s Mars rover (Figure 1) did not come along with a mechanism to be put back into its stable position in case it fell out of it.

Fig 1: A simulation of a wheeled-robot on an unexplored terrain

The legged robots (Figure 2) are perhaps the most intriguing category of robots. Their method of locomotion is extremely complicated, since for every step that they take, they need to raise one limb while maintaining balance on the remaining limbs. This means that the weight of the robot needs to be distributed through an asymmetrically configured body and that always poses as a challenge. Before going towards understanding the balancing in legged robots, steps need to be taken towards figuring out how the robot would move around. This project idea could be summarized as determining the method of locomotion in legged robots. Possible questions that need to be answered in this case are:

  • What is the simplest legged robot design that we can work with? 
  • What kind of sensors do we use to understand how the center of mass shifts with every step?
  • Since the motion here is an actuation of links and joints, how do we determine the kinematics of the robot?

Fig 2: A legged robot going out on a stroll.

There is tremendous scope in this area of understanding balance in robots. Nevertheless, the path to reaching a certain level of expertise needs backing with strong fundamental knowledge in the following topics:

  • CAD Designing Skills
  • Forward and Inverse Kinematics
  • Basics of Electronics and Sensors
  • Computer Programming
  • Machine Learning and Deep Neural Networks 
  • Computer Vision

These are essential for one to be able to venture deeply into robotics and work with its various facets. The project I worked on was an introductory step towards getting a feel for the bigger picture of robotics. The next stage involves answering the questions discussed above and working towards conceptualizing the idea of balancing complex robots.

Authored by: Shivam Kaul
Last edited: May, 2021