Sunteți pe pagina 1din 1

Self-Balancing Robot For Drink Delivery

Yuwei Wang ‘17 and Hung Nguyen ‘17


Department of Engineering
Advisor: Prof. H. Blaise

Introduction Control System Design


Motivated by the annual RoboWaiter competition held by Trinity College An optimal controller is designed using the Linear Quadratic Regulator method. To reduce noise in the accelerometer and to
Engineering Department, this project aims to create a device similar to Sensors including gyroscope, accelerometer and rotary encoders are used for resolve drifting problem with the gyroscope, a
that of an automated drink carrier, but with a major change in design: the real-time state determination. By passing the states through a feedback gain, complimentary filter is used to obtain the tilt
robot is self-balanced on two wheels. With a compact size and good the controller computes the required voltage input to the DC motors to achieve angle of robot
maneuverability, this self-balancing robot is intended for narrow aisles on self-balance. The controller is both simulated and implemented to hardware in
modes of public transportation such as airplanes, trains and buses. The discrete time using Simulink. Finally RC transmitter and receiver are
major challenge lies in the design of an effective controller to achieve implemented to remotely control backward/forward and steering motion of the
self-balance while traveling at a desired speed. robot.
Filtered vs Unfiltered measured angle

Problem Statement
Objective:
To deliver drinks without spilling by using a self-
balancing robot.

Specific Goals:
To stay self-balanced at all time
To be remotely controlled
To be resistant to some degree of disturbance LQR method: weight matrices Q and R are
Arduino Mega2560

determined based on the optimal simulation


Major Components results, and the feedback gain is computed in
MATLAB using Riccati equation

K = [-4.43 -24.08 73.53 5.34]


DC motor& MPU6050 Arduino Mega2560 L298N RC transmitter/
encoder IMU sensor motor driver receiver
Simulation Results Final Design
Mathematical Model Simulation of robot tracking desired states

Dynamic equations of the system are derived and linearized to construct a


state-space model below in terms of state variables and control input.

where

x = displacement of the robot km = motor’s torque constant


ϴ = tilt angle ke = back EMF constant Simulink model for implementation Prototype of the robot
Mp = mass of the robot’s chassis R = terminal resistance
Ip = moment of inertia of the robot’s chassis
l = height of robot’s center of gravity
r = wheel radius Reference & Acknowledgement
Reference
1. David L. Kleinman, P. Krishna Rao, Continuous-Discrete Gain Acknowledgement
Transformation Methods for Linear Feedback Control, 1977 Professor Harry Blaise
2. Yeonhoon Kim, Soo Hyun Kim and Yoon Keun Kwak, Dynamic
Professor John Mertens
Analysis of a Non-holonomic Two-Wheeled Inverted Pendulum, Andrew Musulin
Journal of Intelligent and Robotic Systems, 2005

S-ar putea să vă placă și