Introduction

A lot of innovation is happening in the field of quadrupeds and humanoid robotics, such as quadrupeds like Unitree G2, Boston Dynamics Spot, etc. These robots have a four legged design compared to a traditional four wheeled robot as it allows them to cross uneven terrain such as city streets or forests with large changes in elevation. We are planning to invent a new method of locomotion through a jumping robot design. We are currently looking at different ways of packaging the exterior robot, through a frog leg design or a spherical jumping ball.

Quadruped Traversing uneven environment

China's high tech wheeled defense robot

Problem Statement

The main issue to solve with this robot is creating a design that can store enough potential energy to mimic projectile motion after the jump. We have defined our requirements for our robot to be able to jump over a 1 width foot square box. With attempting a spherical robot or frog robot, we would need to be able to pitch the robot so the velocity vector would have a x component in the positive direction. Creating the stored energy adds complexity to the design as we would need a complex linkage system (ideally following hind legs of a frog) to store the potential energy, and a mechanism that can act as a release and reload. Although the motion of the robot will follow a relatively simple motion profile, the mechanism needed to follow the motion will require complex mechanisms. The problems are summarized below.

Simple Diagram of Requirements

Problem	Example
Launching Mechanism	Hind legs of a frog
Orienting Mechanism	Person orienting slingshot towards target
Loading Mechanism	Stretching a rubber band on a slingshot
Releasing Mechanism	Person letting go of slingshot with hand

Mechanism

Spring Leg Mechanism

Motion Analysis of Design

	Motion Gen Analysis (1 DOF)

	Motion Gen Analysis (1 DOF)
We will be using positional and velocity analysis techniques learned from class to calculate the motion profile of the end effector of the body (assuming is the fixed point). This will dictate how we modify the lengths of the link to map the desired compression In addition, we were thinking of using extension/compression springs with a viscoelastic band to create the spring potential energy	Note that the red slider shown above is meant to be modeled as the ground

Gruebler's Equation for Hind Leg Mechanism w/ Analysis

Spring Release Mechanism

Motion Analysis of Design

*Note the L6 will be controlled by a motor with a worm gear to prevent back driving and enough torque to launch the mechanism	Rough Sketch of crank slider & Ratcheting mechanism

*Note the L6 will be controlled by a motor with a worm gear to prevent back driving and enough torque to launch the mechanism	Rough Sketch of crank slider & Ratcheting mechanism

Gruebler's Equation for Ratcheting and Release Mechanism w/ Analysis

n = 5 (3 gears, 1 coupler link, 1 ground)

J₁ = 5 (4 revolute, 1 prismatic)

J₂ = 1 (1 gear mesh interface)

M = 3* (n−1)−2 * J₁ − J₂

M = 3 * (5-1) - 2 * (5) - 1 = 1

With the addition of a motor attached to the slider gear, we can add an additional DOF

M_tot = M + F actuator

M_tot = 1 +1 = 2 DOF

Enclosure 1 (Spherical)

The weight pendulum would act as a way to shift weight side to side in order to orient the to jump in a certain direction (Additional servo)
Since it is only doing a jump in the box plane, there won’t be weighted pendulums to control the other directions
The robot will be designed to be bottom heavy so it always lands on the foot

Enclosure 2 (Frog) - Backup Design

In the foreleg, the joint between the upper and lower linkage will be a compliant DOF as its only used to support the robot
There will be two hind legs with the spring leg mechanism shown above. They will both use the same ratcheting mechanism to simplify the design

Proposed Scope

Scope: For the final project, we aim to have a robot that can jump over a 1 by 1 by 1 foot box as previously mentioned. Due to complexity associated with the jumping mechanism, we are not planning on using any sensors or adding any autonomous function to the robot, compared to other quadrupeds on the market. We will be using encoders/stepper motions to help control the ratchet mechanism, but will lack other sensors or i/o.

Prior Analysis: We have already identified that we will be using a spring leg mechanism. We would also like to do force and energy calculations to determine the force/energy required to have the robot jump. This would clarify the power needed in motors, springs, and whether the structure of the robot can sustain these forces. To expand on this idea, position, velocity, and force analysis will be conducted.

Future Goals/Additional Steps: We would first need to fabricate both the leg and ratchet mechanism to ensure they output the correct position in practice. We would then need to validate the force output that the leg ratchet mechanism exhibits through spring elastic calculations. A chassis would then be constructed to house these components. Lastly, a lot of experimental and calibration work would be required in order to have the robot successfully jump and land.

Related Interests: We have members that are a part of Dr. Sentis’ Human Centered Robotics Laboratory, where current work is being done on an underactuated robotic hand. We also have members of the group who are in the robotics minor, having taken the Gateway to Robotics course, where we have built & coded Stanford’s pupper robotic leg.

Links Referenced

Yim's jumping robot inspired by squirrels' branch-to-branch leaping technique

Squirrels Inspire Leaping Strategy for Salto Robot

(PDF) DESIGN OF AN ALL-TERRAIN SPHERICAL JUMPING ROBOT WITH HIGH-DYNAMIC MOTION

13.1 -- Project Proposal