Balance

←Back to Primary Functions

Purpose

With the goal of enabling the robot to traverse uneven terrain and travel underwater, subsystems that maintain balance become keystones of the robot’s operational framework.

Design Objective

There are two subsets of the design that influence the implementation of balance control: terrestrial movement and movement in water. Both of these sub-functions vary widely in the possibilities for their design potential, but all possible designs share some common necessities. For terrestrial movement, the robot must be able to recognize what position it is in and where that position is relative to the ground. For movement in water, there are again two subsets of the design requirements namely, movement on the surface of a body of water and movement while submerged. In both situations, the robot must maintain correct orientation and do so efficiently. 

It is possible that many of these functions will be accomplished through the implementation of inertia sensors and actuators: accelerometers, gyroscopes and magnetometers provide feedback on the orientation of the robot and such devices as flywheels, ballasts, and solenoids will assist in correcting the orientation. 

Technical Concepts

  • Underwater Balancing
    • The design of the hull and its balance characteristics are heavily dependent on a variety of factors. Weight and volume distribution determine the center of gravity and center of buoyancy. Some designs allow for moving the center of gravity within the hull, making balance possible without the use of thrusters. The robot will likely need some sort of ballast/air pocket system that is autonomously able to redistribute water/air to properly change its own buoyancy. 
    • Contributions of hull design to drag and imbalance in water
  • Legged Balancing
    • Ground Reaction Force (GRF)
      • One of the designs discussed involves the possibility of a legged robot. Such a setup would require inertial sensors and sensors on the feet/pads to give feedback on the placement of the legs.  The GRF is just what it sounds like: how much force the ground is exerting on the foot, or how much weight the robot is placing on that foot.
    • Center of Pressure (CoP)
      • Combined with the GRF, these allow the robot to deal with different frictional properties at each foot.  The robot determines the desired momenta and optimal foot GRFs and CoPs, allowing for a momentum based stepping algorithm that is more robust than one that only functions via distribution of mass.
  • Wheeled Balancing
    • A robot with wheels does not generally have many ways of actively balancing itself, but it can know what terrain and slopes would cause it to capsize. This can be done with simple calculations of the location of center of gravity and then limiting the robot to not attempting something that it cannot accomplish. Some combination of wheels and legs, or using the proposed grabbing/lifting arm to balance itself, could run on the same concepts as a legged robot.

 

Metrics

  • Ability for robot to travel through a range of depths
  • Robot must prevent capsizing in water or be able to operate regardless of orientation
  • Walking robot must be able to walk at a certain speed across terrain of a certain grade/unevenness without capsizing
  • Robot must be able to maintain a set orientation for operation in certain conditions

Constraints

  • Allocable space within the robot for ballasts
  • Waterproofing rating achievable
  • Ability to implement multiple sensors in creating redundancy
  • Level of precision achievable for a legged robot

Next Steps:

Much of balance as a function depends on what subfunctions are chosen for the robot, the biggest factor here being the choice of legs vs. wheels. As we already have experience with autonomous reaction to stimuli in controls, most of what we need to learn for balance is the math. Once we decide specifically how the robot will move, we can familiarize ourselves with how similar robots balance themselves and implement that in our design.

←Back to Primary Functions

Citations

  1. Chakraborty, S. (2021, September 8). Understanding stability of submarine. Marine Insight. Retrieved September 26, 2021, from https://www.marineinsight.com/naval-architecture/understanding-stability-submarine/
  2. Detweiler C., Sosnowski S., Vasilescu I., Rus D. (2009) Saving Energy with Buoyancy and Balance Control for Underwater Robots with Dynamic Payloads. In: Khatib O., Kumar V., Pappas G.J. (eds) Experimental Robotics. Springer Tracts in Advanced Robotics, vol 54. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-00196-3_49
  3. G. Chen, S. Guo, B. Hou and J. Wang, “Virtual Model Control for Quadruped Robots,” in IEEE Access, vol. 8, pp. 140736-140751, 2020, doi: 10.1109/ACCESS.2020.3013434.
  4. Lee, SH., Goswami, A. A momentum-based balance controller for humanoid robots on non-level and non-stationary ground. Auton Robot 33, 399–414 (2012). https://doi.org/10.1007/s10514-012-9294-z
  5. Raibert, M. H. (2000). Legged robots that balance. MIT Press.