Problem StatementObjectiveAlgorithmsAccelerometer and Gyroscope (Sensor Interface Module)Shaft EncodersMerging the dataChangesProblemsBudget1Critical Design Review Ajay Joseph Saba Rizvi Steve Parkinson Lee Fithian1TABLE OF CONTENTS Problem Statement .....................................................................................................2 Objective ....................................................................................................................2 Algorithms .................................................................................................................3 Accelerometer and Gyroscope (Sensor Interface Module)......................................................... 3 Shaft Encoders ............................................................................................................................ 4 Merging the data ......................................................................................................................... 7 Changes ......................................................................................................................8 Problems.....................................................................................................................9 Budget ......................................................................................................................122 Problem Statement We are to build a mobile robot to follow a pre-specified path using a synthesis of navigation methods. We will execute trial runs and record actual versus specified paths. We will analyze the results to determine the accuracy of the integrated system. Objective Construction: We plan to build a robot consisting of a chassis, power supply, servo motors, and a micro controller. We will use accelerometers, shaft encoders, a gyroscope and a compass as our movement sensors. We will write drivers for the robot using software compatible to the micro controller. We will interface between a personal computer and the robot using serial ports. Experimentation: We will load a path for the robot to follow. We will layout a path on the experiment surface. We will then be able to measure the deviation in the paths. Analysis: The experiment surface will be a grid. We will be able to measure the distance from the target for the different trials. We will then be able to make error distributions for the different navigation methods. From the error distributions, we will be able to create an integrated navigation system which takes into account the error inherent in each sensor. We will then run tests for this integrated system and find its error.3 Algorithms We have separated our program into three different functions, each of which has its own job. Sensor-Interface Module – This module’s job will be to obtain the data from the sensors and convert it so that it can be manipulated. Navigation Module – This module will use the obtained data to calculate what the robot’s next move should be. This is the module responsible for the navigation of the robot. Movement Module – This module is responsible for moving the robot in the desired direction using the motor and wheels. We have come up with Sensor Interface Modules for the accelerometer, gyroscope, and encoders. No matter what sensor we are using, we will have three outputs that the Sensor Interface Module will give to the Navigation Module. Those three output are: 1. Heading – the actual heading relative to the starting position. 2. X component – of the distance traveled 3. Y component – of the distance traveled Accelerometer and Gyroscope (Sensor Interface Module) The accelerometer measures acceleration on clock cycles using T1 and T2. Figure 1: Duty Cycle of Accelerometer The algorithm will repeat every time period, which will be set by us. A probable value is 5 ns. The basic idea will be to reset the counter while the output is low.4 General Formula for Position: Position = Starting Position + (A * t²)/2 Error Formula – To be used to predict error over time Error(Position) = (A * t²)/2 1. Sample axis counters x and y -> store count c = counter value, f = frequency in Hz of clock, n = bits of counter A 8-bit counter gives the following resolution assuming the nominal 12.5% duty cyle. 4/ (2n * .5) Î resolution in g 4/ (28 * .5) Î .03125 g/bit resolution with 8-bit 2. Calculate distance traveled since last sample A = ((T1/T2 - .5))/12.5% Distx = (Ax * t2)/2 (Formula for g based on T1x, T2 is constant) Disty = (Ay * t2)/2 (Formula for g based on T1y, T2 is constant) 3. Add distance to last distance since program started. 4. Sample value of Gyroscope sg = sampled value, ng = nominal value sg – ng / ng *150 = Angular Value = Vang Angle traversed since last sample = Vang (tcurrent – tprevious) = ∆Angle 5. Add angle ∆Angle to Heading. Shaft Encoders Sensor output The output of the shaft encoders is a sine wave. Depending on the resolution, a certain number of pulses are generated for each revolution. Our encoder will generate 256 pulses per5 revolution; half of this resolution is due to the 2 channels of output that are 90° out of phase. The shaft encoder has 5 pins, of which two are used to output the two sine waves. Figure 2: Incremental Encoder Output Connected to the shaft encoder, we will have an encoder to counter interface board. The purpose of this interface will be to provide a counter with a data that it will understand directly from the encoder. It will take as inputs the sine waves generated by the encoder. It will output either an up or down clock. The clock will be used to run the counter. We will have a basic up/down counter to accumulate the number of pulses generated by the encoder. The input will be from the encoder/counter interface and provide the clock for the counter. The counter will then increment as the wheel rotates. When the counter is full, an interrupt will be generated and the information that the maximum value of the counter number of rotations have occurred. From the output of the counter we will be able to determine the robot’s position.6 The position algorithm has been determined before and the equations are well known1. The basis for the algorithm is counting the number of revolutions and translating that into distance. The data produced by the counter will be the number or fraction of wheel revolutions. When the counter is full, a certain amount of distance will have been traveled.