我正在为我的Arduino编写程序，以在图表上绘制用户输入的数据，并给出最合适的直线+直线方程，回归值和确定的系数（ R ^ 2 ）值。 Arduino的串行监视器无法完成打印整个输出。有什么问题，我该如何解决？
我已经读到这可能是内存问题，因此我对程序中文字字符串f() macros的所有实例使用Serial.prints大大减少了动态内存的使用量。这有助于显着减少正在使用的内存量，但最终串行监视器无法完成整个输出的打印。此外，我知道较大的Arduino可能会更好地工作，但由于该项目的限制，我需要使用Arduino Uno模型，因此无法更改。
我希望串行监视器接收用户的数据，然后打印出最佳拟合线，回归误差，计算出的Y values和确定值系数的方程式。相反，串行监视器仅打印最适合行方程的前几个字符，然后突然停止打印。
#include <math.h>
//#include <EEPROM.h>
#define ARDBUFFER 16
#include <stdarg.h>
#include <Arduino.h>
const byte buffSize = 16;
char inputSeveral[buffSize]; // schar array for input function below
// space for 16 chars and a terminator
byte maxChars = 12; // a shorter limit to make it easier to see what happens
// if too many chars are entered
float* px; // dynamic array for x's (DAQ system values)
float* py; // dynamic array for y's (calibrated values)
// NOTE: DELAYS TEMPORARY - WHILE LOOPS FOR INPUT NOT WORKING
// NOTE: USING 86% MEMORY ON ARDUINO UNO
void setup() {
Serial.begin(9600);
Serial.println(F("Starting..."));
// Menu select for function to fit against
Serial.println(F("Select fit: "));
Serial.println(F(" (1)Linear - Minimum two points"));
Serial.println(F(" (2)Quadratic - Minimum three points"));
Serial.println(F(" (3)Exponential - Minimum three points, y != 0")); // float check restrictions on exp, log, power
Serial.println(F(" (4)Logarithmic - Minimum three points, x != 0"));
Serial.println(F(" (5)Power - Minimum three points, x != 0"));
Serial.println(F(" (0)Exit"));
delay(3000);
readSeveralChars();
uint8_t fitChoice = atoi(inputSeveral);
// Exit
if (fitChoice == 0)
{
Serial.print(F("Exiting calibration process..."));
delay(2000);
exit(0);
}
// Linear
if(fitChoice == 1) {
Serial.println(F("Fit Chosen: Linear"));
Serial.print(F("Input total points: ")); // prompt user
delay(2000); // delay for input (while (Serial.avaiable()) causes char array to become zero and instantly changes input variable to 0)
readSeveralChars();
uint8_t totalPoints = atoi(inputSeveral); // converts char array to int
Serial.println(totalPoints);
Serial.println(F("NOTE - X's are DAQ system values measured, Y's are final unit calibrated values"));
// Error and warning checks for minimum points
if (totalPoints < 2)
{
Serial.print(F("At least two points needed for linear. Restarting calibration process..."));
delay(2000);
setup();
}
else if (totalPoints == 2)
{
Serial.println(F("WARNING - Minimum points met. Overdefined recommended."));
}
delay(3000);
px = new float[totalPoints]; // Load x's into array
py = new float[totalPoints]; // Load y's into array
float rCoeff;
for (uint8_t i = 0; i < totalPoints; ++i) // loop through arrays and fill in values by input
{
ardprintf("Input x%d", i+1); // printf for serial, function implemented below
delay(2000);
readSeveralChars();
px[i] = atof(inputSeveral);
Serial.println(px[i]);
delay(1000);
ardprintf("Input y%d", i+1);
delay(2000);
readSeveralChars();
py[i] = atof(inputSeveral);
Serial.println(py[i]);
delay(1000);
}
fabls_linear(totalPoints, px, py); // send inputed points to fabls calculator
rCoeff = determinationCoefficient(totalPoints, px, py);
Serial.print(F("r^2 = "));
Serial.println(rCoeff);
}
// Quadratic
else if (fitChoice == 2) {
Serial.println(F("Fit Chosen: Quadratic"));
Serial.print(F("Input total points: "));
delay(2000);
readSeveralChars();
uint8_t totalPoints = atoi(inputSeveral);
Serial.println(totalPoints);
if (totalPoints < 3)
{
Serial.print(F("At least three points needed for quadratic. Restarting calibration process..."));
delay(2000);
setup();
}
else if (totalPoints == 3)
{
Serial.println(F("WARNING - Minimum points met. Overdefined recommended."));
}
delay(3000);
px = new float[totalPoints];
py = new float[totalPoints];
float rCoeff;
for (uint8_t i = 0; i < totalPoints; ++i)
{
ardprintf("Input x%d", i+1);
delay(2000);
readSeveralChars();
px[i] = atof(inputSeveral);
Serial.println(px[i]);
delay(1000);
ardprintf("Input y%d", i+1);
delay(2000);
readSeveralChars();
py[i] = atof(inputSeveral);
Serial.println(py[i]);
delay(1000);
}
fabls_quad(totalPoints, px, py);
rCoeff = determinationCoefficient(totalPoints, px, py);
Serial.print(F("r^2 = "));
Serial.println(rCoeff);
}
// Exponential
else if (fitChoice == 3) {
Serial.println(F("Fit Chosen: Exponential"));
Serial.print(F("Input total points: "));
delay(2000);
readSeveralChars();
uint8_t totalPoints = atoi(inputSeveral);
Serial.println(totalPoints);
if (totalPoints < 3)
{
Serial.print(F("ERROR - At least three points needed for exponential. Restarting calibration process..."));
delay(2000);
setup();
}
else if (totalPoints == 3)
{
Serial.println(F("WARNING - Minimum points met. Overdefined recommended."));
}
delay(3000);
px = new float[totalPoints];
py = new float[totalPoints];
float rCoeff;
for (uint8_t i = 0; i < totalPoints; ++i)
{
ardprintf("Input x%d", i+1);
delay(2000);
readSeveralChars();
px[i] = atof(inputSeveral);
Serial.println(px[i]);
delay(1000);
ardprintf("Input y%d", i+1);
delay(2000);
readSeveralChars();
py[i] = atof(inputSeveral);
Serial.println(py[i]);
if (py[i] == 0) // Catch zero point errors
{
Serial.println(F("ERROR - y's cannot be zero for exponential. Restarting calibration process... "));
delay(2000);
setup();
}
delay(1000);
}
fabls_exp(totalPoints, px, py);
rCoeff = determinationCoefficient(totalPoints, px, py);
Serial.print(F("r^2 = "));
Serial.println(rCoeff);
}
// Logarithmic
else if (fitChoice == 4) {
Serial.println(F("Fit Chosen: Logarithmic"));
Serial.print(F("Input total points: "));
delay(2000);
readSeveralChars();
uint8_t totalPoints = atoi(inputSeveral);
Serial.println(totalPoints);
if (totalPoints < 3)
{
Serial.print(F("At least three points needed for logarithmic. Restarting calibration process..."));
delay(2000);
setup();
}
else if (totalPoints == 3)
{
Serial.println(F("WARNING - Minimum points met. Overdefined recommended."));
}
delay(3000);
px = new float[totalPoints];
py = new float[totalPoints];
float rCoeff;
for (uint8_t i = 0; i < totalPoints; ++i)
{
ardprintf("Input x%d", i+1);
delay(2000);
readSeveralChars();
px[i] = atof(inputSeveral);
Serial.println(px[i]);
if (px[i] == 0)
{
Serial.println(F("ERROR - x's cannot be zero for logarthimic. Restarting calibration process..."));
delay(2000);
setup();
}
delay(1000);
ardprintf("Input y%d", i+1);
delay(2000);
readSeveralChars();
py[i] = atof(inputSeveral);
Serial.println(py[i]);
delay(1000);
}
fabls_log(totalPoints, px, py);
rCoeff = determinationCoefficient(totalPoints, px, py);
Serial.print(F("r^2 = "));
Serial.println(rCoeff);
}
// Power
else if (fitChoice == 5) {
Serial.println(F("Fit Chosen: Power"));
Serial.print(F("Input total points: "));
delay(2000);
readSeveralChars();
uint8_t totalPoints = atoi(inputSeveral);
Serial.println(totalPoints);
if (totalPoints < 3)
{
Serial.print(F("At least three points needed for power. Restarting calibration process..."));
delay(2000);
setup();
}
else if (totalPoints == 3)
{
Serial.println(F("WARNING - Minimum points met. Overdefined recommended."));
}
delay(3000);
px = new float[totalPoints];
py = new float[totalPoints];
float rCoeff;
for (uint8_t i = 0; i < totalPoints; ++i)
{
ardprintf("Input x%d", i+1);
delay(2000);
readSeveralChars();
px[i] = atof(inputSeveral);
Serial.println(px[i]);
if (px[i] == 0)
{
Serial.println(F("ERROR - x's cannot be zero for power. Restarting calibration process..."));
delay(2000);
setup();
}
delay(1000);
ardprintf("Input y%d", i+1);
delay(2000);
readSeveralChars();
py[i] = atof(inputSeveral);
Serial.println(py[i]);
delay(1000);
}
fabls_power(totalPoints, px, py);
rCoeff = determinationCoefficient(totalPoints, px, py);
Serial.print(F("r^2 = "));
Serial.println(rCoeff);
}
// Invalid
else {
Serial.println(F("Invalid choice. Restarting calibration process..."));
delay(2000);
setup(); // Restart, jumps backs to beginning
}
// deallocation (move to after EEPROM)
delete[] px;
delete[] py;
// LOAD EEPROM
// Once input points are given and regression data is returned
// prompt user to send new calibration values to EEPROM
}
void loop() {
// None
}
// https://forum.arduino.cc/index.php?topic=96292.0
// Fit Analysis By Least Squares
float alog(float x)
{ return (x < 0) ? -log(-x) : ((x > 0) ? log(x) : 0);
}
void fabls_linear(unsigned int n,float *px,float *py)
{
float regressedValueArray[buffSize];
byte mask='\x00',sign,sign2;
unsigned int i;
int least=-1;
float beta,d2,denom,dy,p,percent_error,r=(n-1),x,y,yc;
float a1,a2,a3,s,s1,s2,s3,s4,s5,s6,s7,z[5];
byte *f = "%f %f %f %f %f\n";
s1 = s2 = s3 = s4 = s = 0;
for (i=0; i<n; i++)
{ x = px[i];
y = py[i];
s1 += x;
s2 += x * x;
s3 += y;
s4 += x * y;
}
if (denom = n * s2 - s1 * s1)
{ a1 = (s3 * s2 - s1 * s4) / denom;
a2 = (n * s4 - s3 * s1) / denom;
for (i=0; i<n; i++)
{ dy = py[i] - (a2 * px[i] + a1);
s += dy * dy;
}
s = sqrt(s / r);
sign = (a1 < 0) ? '-' : '+';
Serial.print(F("CHECKPOINT #1"));
ardprintf("Linear: y = (%f) x %c %f; s = %f\n",a2,sign,fabs(a1),s);
Serial.print(F("CHECKPOINT #2"));
mask |= '\x01';
z[0] = s;
}
Serial.print(F("X"));
Serial.print(F(" Y"));
Serial.print(F(" Calculated Y"));
Serial.println(F(" PercentError%"));
for (unsigned int i = 0; i < n; ++i)
{
float y = (a2) * px[i] + (a1);
// PercentError%=((regressionvalue-calibrationvalue)/calibrationvalue)*100
regressedValueArray[i] = y;
float error = ((y - py[i])/py[i])*100;
ardprintf("%f %f %f %f", px[i], py[i], y, error);
}
determinationCoefficient(n, py, regressedValueArray);
}
void fabls_quad(unsigned int n,float *px,float *py)
{
float regressedValueArray[buffSize];
byte mask='\x00',sign,sign2;
unsigned int i;
int least=-1;
float beta,d2,denom,dy,p,percent_error,r=(n-1),x,y,yc;
float a1,a2,a3,s,s1,s2,s3,s4,s5,s6,s7,z[5];
byte *f = "%f %f %f %f %f\n";
s1 = s2 = s3 = s4 = s5 = s6 = s7 = s = 0;
for (i=0; i<n; i++)
{ x = px[i];
y = py[i];
s1 += x;
s2 += x * x;
s3 += x * x * x;
s4 += x * x * x * x;
s5 += y;
s6 += x * y;
s7 += x * x * y;
}
if (denom = n * (s2 * s4 - s3 * s3) -
s1 * (s1 * s4 - s2 * s3) +
s2 * (s1 * s3 - s2 * s2) )
{ a1 = (s5 * (s2 * s4 - s3 * s3) -
s6 * (s1 * s4 - s2 * s3) +
s7 * (s1 * s3 - s2 * s2)) / denom;
a2 = (n * (s6 * s4 - s3 * s7) -
s1 * (s5 * s4 - s7 * s2) +
s2 * (s5 * s3 - s6 * s2)) / denom;
a3 = (n * (s2 * s7 - s6 * s3) -
s1 * (s1 * s7 - s5 * s3) +
s2 * (s1 * s6 - s5 * s2)) / denom;
for (i=0; i<n; i++)
{ x = px[i];
dy = py[i] - (a3 * x * x + a2 * x + a1);
s += dy * dy;
}
s = sqrt(s / r);
sign = (a1 < 0) ? '-' : '+';
sign2 = (a2 < 0) ? '-' : '+';
ardprintf("Quadratic: y = (%f) x^2 %c (%f) x %c %f; s = %f\n",
a3,sign2,fabs(a2),sign,fabs(a1),s);
mask |= '\x02';
z[1] = s;
}
Serial.print(F("X"));
Serial.print(F(" Y"));
Serial.print(F(" Calculated Y"));
Serial.println(F(" PercentError%"));
for (unsigned int i = 0; i < n; ++i)
{
float y = (a2) * px[i] + (a1);
// PercentError%=((regressionvalue-calibrationvalue)/calibrationvalue)*100
regressedValueArray[i] = y;
float error = ((y - py[i])/py[i])*100;
ardprintf("%f %f %f %f", px[i], py[i], y, error);
}
determinationCoefficient(n, py, regressedValueArray);
}
void fabls_exp(unsigned int n,float *px,float *py)
{
float regressedValueArray[buffSize];
byte mask='\x00',sign,sign2;
unsigned int i;
int least=-1;
float beta,d2,denom,dy,p,percent_error,r=(n-1),x,y,yc;
float a1,a2,a3,s,s1,s2,s3,s4,s5,s6,s7,z[5];
byte *f = "%f %f %f %f %f\n";
s1 = s2 = s3 = s4 = s = 0;
for (i=0; i<n; i++)
{ x = px[i];
y = alog(py[i]);
s1 += x;
s2 += x * x;
s3 += y;
s4 += x * y;
}
if (denom = n * s2 - s1 * s1)
{ a1 = (s3 * s2 - s1 * s4) / denom;
a2 = (n * s4 - s3 * s1) / denom;
for (i=0; i<n; i++)
{ dy = alog(py[i]) - (a2 * px[i] + a1);
s += dy * dy;
}
s = sqrt(s / r);
sign = (a1 < 0) ? '-' : '+';
ardprintf("Exponential: y = exp(%f x %c %f); s = %f\n",a2,sign,fabs(a1),s);
mask |= '\x04';
z[2] = s;
}
Serial.print(F("X"));
Serial.print(F(" Y"));
Serial.print(F(" Calculated Y"));
Serial.println(F(" PercentError%"));
for (unsigned int i = 0; i < n; ++i)
{
float y = (a2) * px[i] + (a1);
// PercentError%=((regressionvalue-calibrationvalue)/calibrationvalue)*100
regressedValueArray[i] = y;
float error = ((y - py[i])/py[i])*100;
ardprintf("%f %f %f %f", px[i], py[i], y, error);
}
determinationCoefficient(n, py, regressedValueArray);
}
void fabls_log(unsigned int n,float *px,float *py)
{
float regressedValueArray[buffSize];
byte mask='\x00',sign,sign2;
unsigned int i;
int least=-1;
float beta,d2,denom,dy,p,percent_error,r=(n-1),x,y,yc;
float a1,a2,a3,s,s1,s2,s3,s4,s5,s6,s7,z[5];
byte *f = "%f %f %f %f %f\n";
s1 = s2 = s3 = s4 = s = 0;
for (i=0; i<n; i++)
{ x = alog(px[i]);
y = py[i];
s1 += x;
s2 += x * x;
s3 += y;
s4 += x * y;
}
if (denom = n * s2 - s1 * s1)
{ a1 = (s3 * s2 - s1 * s4) / denom;
a2 = (n * s4 - s3 * s1) / denom;
for (i=0; i<n; i++)
{ x = alog(px[i]);
dy = py[i] - (x * a2 + a1);
s += dy * dy;
}
s = sqrt(s / r);
sign = (a1 < 0) ? '-' : '+';
ardprintf("Logarithmic: y = (%f) ln(x) %c %f; s = %f\n",a2,sign,fabs(a1),s);
mask |= '\x08';
z[3] = s;
}
Serial.print(F("X"));
Serial.print(F(" Y"));
Serial.print(F(" Calculated Y"));
Serial.println(F(" PercentError%"));
for (unsigned int i = 0; i < n; ++i)
{
float y = (a2) * px[i] + (a1);
// PercentError%=((regressionvalue-calibrationvalue)/calibrationvalue)*100
regressedValueArray[i] = y;
float error = ((y - py[i])/py[i])*100;
ardprintf("%f %f %f %f", px[i], py[i], y, error);
}
determinationCoefficient(n, py, regressedValueArray);
}
void fabls_power(unsigned int n,float *px,float *py)
{
float regressedValueArray[buffSize];
byte mask='\x00',sign,sign2;
unsigned int i;
int least=-1;
float beta,d2,denom,dy,p,percent_error,r=(n-1),x,y,yc;
float a1,a2,a3,s,s1,s2,s3,s4,s5,s6,s7,z[5];
byte *f = "%f %f %f %f %f\n";
s1 = s2 = s3 = s4 = s = 0;
for (i=0; i<n; i++)
{ x = alog(px[i]);
y = alog(py[i]);
s1 += x;
s2 += x * x;
s3 += y;
s4 += x * y;
}
if (denom = n * s2 - s1 * s1)
{ a1 = exp((s3 * s2 - s1 * s4) / denom);
a2 = (n * s4 - s3 * s1) / denom;
for (i=0; i<n; i++)
{ dy = py[i] - a1 * pow(px[i],a2);
s += dy * dy;
}
s = sqrt(s / r);
sign = (a1 < 0) ? '-' : '+';
ardprintf("Power: y = (%f) x ^ (%f); s = %f\n",a1,a2,s);
mask |= '\x10';
z[4] = s;
}
Serial.print(F("X"));
Serial.print(F(" Y"));
Serial.print(F(" Calculated Y"));
Serial.println(F(" PercentError%"));
for (unsigned int i = 0; i < n; ++i)
{
float y = (a2) * px[i] + (a1);
// PercentError%=((regressionvalue-calibrationvalue)/calibrationvalue)*100
regressedValueArray[i] = y;
float error = ((y - py[i])/py[i])*100;
ardprintf("%f %f %f %f", px[i], py[i], y, error);
}
determinationCoefficient(n, py, regressedValueArray);
}
void readSeveralChars() {
// this reads all the characters in the input buffer
// if there are too many for the inputSeveral array the extra chars will be lost
inputSeveral[0] = 0; // makes inputSeveral an empty string with just a terminator
byte ndx = 0; // the index position for storing the character
if (Serial.available() > 0) {
while (Serial.available() > 0) { // keep going until buffer is empty
if (ndx > maxChars - 1) { // -1 because arrays count from 0
ndx = maxChars; // if there are too many chars the extra ones are
} // dumped into the last array element which will
// be overwritten by the string terminator
inputSeveral[ndx] = Serial.read();
ndx ++;
}
if (ndx > maxChars) { // to make sure the terminator is not written beyond the array
ndx = maxChars;
}
inputSeveral[ndx] = 0; // add a zero terminator to mark the end of the string
}
}
// https://gist.github.com/asheeshr/9004783
// A printf function for serial communication from Arduino boards
int ardprintf(char *str, ...)
{
int i, count=0, j=0, flag=0;
char temp[ARDBUFFER+1];
for(i=0; str[i]!='\0';i++) if(str[i]=='%') count++;
va_list argv;
va_start(argv, count);
for(i=0,j=0; str[i]!='\0';i++)
{
if(str[i]=='%')
{
temp[j] = '\0';
Serial.print(temp);
j=0;
temp[0] = '\0';
switch(str[++i])
{
case 'd': Serial.print(va_arg(argv, int));
break;
case 'l': Serial.print(va_arg(argv, long));
break;
case 'f': Serial.print(va_arg(argv, float));
break;
case 'c': Serial.print((char)va_arg(argv, int));
break;
case 's': Serial.print(va_arg(argv, char *));
break;
default: ;
};
}
else
{
temp[j] = str[i];
j = (j+1)%ARDBUFFER;
if(j==0)
{
temp[ARDBUFFER] = '\0';
Serial.print(temp);
temp[0]='\0';
}
}
};
Serial.println();
return count + 1;
}
// Extra goodness of fit information
float determinationCoefficient(int n, float *y, float *yRegression)
{
float averageY = 0.0f;
float squareDiffSumY = 0.0f;
float regressDifferenceY = 0.0f;
float regressDiffSumY = 0.0f;
for (int i = 0; i < n; i++)
{
averageY += y[i];
}
averageY /= n;
for (int i = 0; i < n; i++)
{
squareDiffSumY += ((y[i] - averageY) * (y[i] - averageY));
}
for (int i = 0; i < n; i++)
{
regressDiffSumY += ((yRegression[i] - averageY) * (yRegression[i] - averageY));
}
return regressDiffSumY / squareDiffSumY;
}
答案 0 :(得分：0)
首先，检查绑定速率是否在您的代码中并且串行监视器是否相等。然后运行用于您的打印订单的检查代码。有时，运行时错误或逻辑错误会导致代码无法执行。