SPRING 2021 MAE 156B SPONSORED PROJECT
UNIVERSITY OF CALIFORNIA, SAN DIEGO
SPONSORED BY DAVID GREGG, ATA Engineering, Inc.
Figure 1: SP21 Project Final Picture
Figure 2: SP19 Animatronic Figure shown lifting a cup
(Credit: 2019 Team Final Report).
Figure 3: SP20 Animatronic Figure shown performing a trapezoidal motion profile (left) vs polynomial motion profile (right).
(Credit: 2020 Team Website)
Background
Our sponsor, ATA Engineering, is an engineering consultant services company headquartered in San Diego that specializes in the design, testing, and analysis of highly engineered structures subject to severe dynamic loads for a wide range of industries. One area of specialty is the entertainment industry. Robots used by this industry in mechatronic applications are commonly referred to as animatronic robotic figures. These animatronic figures are commonly employed in the entertainment industry as robotic puppets meant to portray characters. For the past two years, the company has collaborated with UCSD design teams to improve the testing and analysis capabilities for animatronic figures, leading to the development of our team’s animatronic figures, with an actuated arm and torso. The 2019 project team designed and fabricated the initial animatronic figures with an actuated arm, torso, and preliminary control system. The 2020 project team refined and updated the control system to reduce vibration and produced a preliminary design for a breakaway joint. The breakaway joint is a device placed inline at the upper arm section which will mechanically disconnect when subjected to forces or moments which exceed its specifically set mechanical parameters (breakaway force or moment). The sponsor desired that these parameters be tunable and that the joint be resettable so the breakaway event can be repeated as necessary. The current animatronic figure was provided to the 2021 team functional and controllable using the Arduino control system, but required further development to meet the sponsor’s desire.
Objectives
The torso turntable should be able to support at maximum a 20 Nm tilting moment from the arm mechanism and rotate about its axis without wobbling.
The breakaway joint should be able to provide different levels of force for break conditions.
The force levels should be adjustable without removing the link and with a minimal reconfiguration time. 5 minutes is acceptable.
The breakaway joint should include some mechanism to prevent damage on the arm when breaking.
The analytical control system to define a motion profile that will achieve the intended goals will not be an active control system that uses a feedback loop. It will be a predictive control system that will define motion profiles based on desired force levels in the joint, and then the defined motion profile will be run with the robotic figure.
The accuracy of the analytical motion profile will require the model to be correlated with physical test data so that physical characteristics of the figure, such as motor function characteristics, are accounted for in the control algorithm.
Minimize the weight of the arm.
Control flexibility in range of motion profiles.
The arm can break into multiple directions.
Design that can breakaway in four directions (up, down, left, right).
Simplified design to achieve objectives without added complexity.
Perform dynamic modeling of the figure and find a transmissibility function.
Infinite controllability of the force settings.
Final Design
Figure 5: Final Picture of the New Robot
For the final design, a new breakaway joint using electromagnet was added, a new slewing bearing on the torso-turntable was replaced, and the control system has been updated.
Figure 6: Flexmag link
Change to one side electromagnet for different magnitudes of pull force. The other side is metal plate to maximize the magnet performance.
Figure 7: Electromagnet Breakaway Joint
Figure 8: The Old Torso-Turntable
Replace turntable with slewing bearing which can withstand tilting moment
Figure 9: The New Slewing Bearing Table
Figure 8: Original control hardware setup
Using Arduino and Dynamixel Shield, different motion profiles can be input to motors by MATLAB.
Besides the system, a control box is built to put all hardware inside. the magnet power can also be manually controlled by a knob on the box.
Figure 9: Updated control hardware setup
Performance Results
Static Test
Breakaway moment from static analysis is 6.41Nm, assuming the motor torque is at stall torque of 10Nm (Breakaway Moment = Weight of Arm x Distance to Breakaway Joint - Motor Torque)
Breakaway moment from simulation is 5.2 Nm for the arm held out fully extended, which is a percent error of 18.9% compared to breakaway moment from static analysis
The two values are comparable, which means the dynamic model is sufficiently accurate
Dynamic Profiles Running
Breakaway moment from dynamic analysis is 5Nm compared to 6.41Nm from static analysis, which is comparable (18.9% difference)
Difference between the two cases (for dynamic analysis) are the rate at which the moment increases due to the differing test conditions (low vs high speeds and acceleration)
Breakaway moments generated for high accelerations using dynamic analysis will be more accurate for our final design objective to predict arm breakage
Dynamic Test
Executive Summary