Autonomous Robot that navigates using ultrasonic sensor and an artificial neural network to change directions
Neural networked tested with XOR truth table, then trained with distances
example outputs from NN:
Arduino NN XOR outputs:
Training Pattern: 0
Input 1 1 Target 0 Output 0.00735
Training Pattern: 1
Input 0 1 Target 1 Output 0.99400
Training Pattern: 2
Input 0 0 Target 0 Output 0.00530
Training Pattern: 3
Input 1 0 Target 1 Output 0.99399
Training Pattern: 4
Input 0 0 Target 0 Output 0.00530
Training Pattern: 5
Input 1 0 Target 1 Output 0.99399
Training Pattern: 6
Input 0 1 Target 1 Output 0.99400
Training Pattern: 7
Input 1 1 Target 0 Output 0.00735
Training Pattern: 8
Input 0 0 Target 0 Output 0.00530
Training Pattern: 9
Input 0 1 Target 1 Output 0.99400
Training Pattern: 10
Input 1 0 Target 1 Output 0.99399
Training Pattern: 11
Input 0 0 Target 0 Output 0.00530
Training Pattern: 12
Input 1 1 Target 0 Output 0.00735
Training Pattern: 13
Input 0 0 Target 0 Output 0.00530
Training Pattern: 14
Input 0 1 Target 1 Output 0.99400
Training Pattern: 15
Input 1 0 Target 1 Output 0.99399
Training Pattern: 16
Input 1 1 Target 0 Output 0.00735
Training Pattern: 17
Input 0 1 Target 1 Output 0.99400
Training Pattern: 18
Input 1 0 Target 1 Output 0.99399
Training Pattern: 19
Input 0 0 Target 0 Output 0.00530
Training Pattern: 20
Input 1 1 Target 0 Output 0.00735
trained on distances:
Training Pattern: 0
Input 74.0000000000 112.0000000000 Target 1 Output 1.00000
Training Pattern: 1
Input 103.0000000000 36.0000000000 Target 0 Output 0.00000
Training Pattern: 2
Input 80.0000000000 73.0000000000 Target 0 Output 0.23420
Training Pattern: 3
Input 52.0000000000 84.0000000000 Target 1 Output 1.00000
Training Pattern: 4
Input 100.0000000000 31.0000000000 Target 0 Output 0.00000
Training Pattern: 5
Input 68.0000000000 44.0000000000 Target 0 Output 0.00003
Training Pattern: 6
Input 89.0000000000 90.0000000000 Target 1 Output 0.97848
Training Pattern: 7
Input 59.0000000000 25.000000000 Target 0 Output 0.00000
Training Pattern: 8
Input 97.0000000000 104.0000000000 Target 1 Output 0.99939
Training Pattern: 9
Input 24.0000000000 13.0000000000 Target 0 Output 0.01996
Training Pattern: 10
Input 24.0000000000 94.0000000000 Target 1 Output 1.00000
Training Pattern: 11
Input 88.0000000000 91.0000000000 Target 1 Output 0.99345
Training Pattern: 12
Input 94.0000000000 39.0000000000 Target 0 Output 0.00000
Training Pattern: 13
Input 62.0000000000 26.0000000000 Target 0 Output 0.00000
Training Pattern: 14
Input 91.0000000000 67.0000000000 Target 0 Output 0.00004
Training Pattern: 15
Input 105.0000000000 27.0000000000 Target 0 Output 0.00000
Training Pattern: 16
Input 85.0000000000 58.0000000000 Target 0 Output 0.00001
Training Pattern: 17
Input 22.0000000000 37.0000000000 Target 1 Output 0.99999
Training Pattern: 18
Input 65.0000000000 43.0000000000 Target 0 Output 0.00008
Training Pattern: 19
Input 78.0000000000 65.0000000000 Target 0 Output 0.00846
Training Pattern: 20
Input 47.0000000000 40.0000000000 Target 0 Output 0.20267
Training Pattern: 21
Input 55.0000000000 45.0000000000 Target 0 Output 0.04156
Training Pattern: 22
Input 60.0000000000 73.0000000000 Target 1 Output 0.99997
Training Pattern: 23
Input 54.0000000000 90.0000000000 Target 1 Output 1.00000
Training Pattern: 24
Input 78.0000000000 68.0000000000 Target 0 Output 0.04683
Training Pattern: 25
Input 12.0000000000 40.0000000000 Target 1 Output 1.00000
Training Pattern: 26
Input 5.0000000000 54.0000000000 Target 1 Output 1.00000
Training Pattern: 27
Input 7.0000000000 59.0000000000 Target 1 Output 1.00000
Training Pattern: 28
Input 46.0000000000 47.0000000000 Target 1 Output 0.97279
Training Pattern: 29
Input 39.0000000000 10.0000000000 Target 0 Output 0.00000
Training Pattern: 30
Input 24.0000000000 36.0000000000 Target 1 Output 0.99994
Training Pattern: 31
Input 83.0000000000 59.0000000000 Target 0 Output 0.00003
Training Pattern: 32
Input 94.0000000000 102.0000000000 Target 1 Output 0.99964
Training Pattern: 33
Input 6.0000000000 10.0000000000 Target 1 Output 0.99436
Training Pattern: 34
Input 0.0000000000 40.0000000000 Target 1 Output 1.00000
Training Pattern: 35
Input 16.0000000000 69.0000000000 Target 1 Output 1.00000
Training Pattern: 36
Input 43.0000000000 55.0000000000 Target 1 Output 0.99995
Training Pattern: 37
Input 64.0000000000 31.0000000000 Target 0 Output 0.00000
Training Pattern: 38
Input 86.0000000000 90.0000000000 Target 1 Output 0.99634
Training Pattern: 39
Input 20.0000000000 23.0000000000 Target 1 Output 0.99052 87
actual outputs from getResults function:
//inputs
43 //left
143 //right
//output
1.00 // turn right
//inputs
3 // left
42 // right
//output
1.00 // turn right
//inputs
40 //left
42 //right
//output
0.98 // turn right
//inputs
40 // left
10 //right
//output
0.00 // turn left
resources used to write neural network:
http://robotics.hobbizine.com/arduinoann.html
https://www.cs.bham.ac.uk//~jxb/INC/nn.html
https://www.the-diy-life.com/running-an-artificial-neural-network-on-an-arduino-uno/
https://www.coursera.org/learn/neural-networks-deep-learning