Filter: examples: TransferFunctionCheck: add support for varable DT low pass

libraries/Filter/examples/TransferFunctionCheck/FilterTest.m

@@ -5,7 +5,7 @@ clc
 % File generated by the TransferFunctionCheck example
 filename = "test.csv";
 
-% There are some anoying warnings for text parsing and matrix solve
+% There are some annoying warnings for text parsing and matrix solve
 warning('off','MATLAB:rankDeficientMatrix');
 warning('off','MATLAB:table:ModifiedAndSavedVarnames');
 
@@ -21,6 +21,10 @@ fprintf("%s\n",lines)
 
 % Try and work out types
 filters = {};
+filter_index = find(startsWith(lines,"LowPassFilterConstDtFloat"));
+for i = 1:numel(filter_index)
+    filters{end+1} = parse_low_pass(lines, filter_index(i)); %#ok<SAGROW> 
+end
 filter_index = find(startsWith(lines,"LowPassFilterFloat"));
 for i = 1:numel(filter_index)
     filters{end+1} = parse_low_pass(lines, filter_index(i)); %#ok<SAGROW> 
@@ -102,7 +106,7 @@ for i = 1:num_sweep
 
 end
 
-% Caculalte using transfer function
+% Calculate using transfer function
 Z = exp(1i*pi*(freq/(sample_freq*0.5)));
 
 for i = numel(filters):-1:1
@@ -113,7 +117,7 @@ end
 % multiply all transfer functions together
 H_all = prod(H,2); 
 
-% Extract anmpitude and phase from transfer function
+% Extract amplitude and phase from transfer function
 calc_amplitude = abs(H_all);
 calc_phase = atan2d(imag(H_all),real(H_all));
 
@@ -142,8 +146,9 @@ xlabel('Frequency (Hz)')
 ylabel('phase')
 
 function ret = parse_low_pass(lines, start_index)
-    if ~startsWith(lines(start_index),"LowPassFilterFloat")
+    found = startsWith(lines(start_index),"LowPassFilterFloat") || startsWith(lines(start_index),"LowPassFilterConstDtFloat");
-        error("Expecting LowPassFilterFloat")
+    if ~found
+        error("Expecting LowPassFilterConstDtFloat or LowPassFilterFloat")
     end
     
     % Next line gives sample rate and cutoff

libraries/Filter/examples/TransferFunctionCheck/TransferFunctionCheck.cpp

@@ -16,7 +16,7 @@ const AP_HAL::HAL& hal = AP_HAL::get_HAL();
 
 // Some helper classes to allow accessing protected variables, also useful for adding type specific prints
 
-class LowPassHelper : public LowPassFilterConstDtFloat {
+class LowPassConstDtHelper : public LowPassFilterConstDtFloat {
 public:
     using LowPassFilterConstDtFloat::LowPassFilterConstDtFloat;
 
@@ -39,6 +39,33 @@ private:
     float _sample_freq;
 };
 
+class LowPassHelper : public LowPassFilterFloat {
+public:
+    using LowPassFilterFloat::LowPassFilterFloat;
+
+    void set_cutoff_frequency_override(float sample_freq, float new_cutoff_freq) {
+        // Stash the DT so we can use it later
+        _DT = 1.0 / sample_freq;
+        set_cutoff_frequency(new_cutoff_freq);
+    }
+
+    // Were really cheating here and using the same method as the filter to get the coefficient
+    // rather than pulling the coefficient directly
+    void print_transfer_function() {
+        hal.console->printf("LowPassFilterFloat\n");
+        hal.console->printf("Sample rate: %.9f Hz, Cutoff: %.9f Hz\n", 1.0 / _DT, get_cutoff_freq());
+        hal.console->printf("Low pass filter in the form: H(z) = a/(1-(1-a)*z^-1)\n");
+        hal.console->printf("a: %.9f\n", calc_lowpass_alpha_dt(_DT, get_cutoff_freq()));
+    }
+
+    float apply_override(const float sample) {
+        return apply(sample, _DT);
+    }
+
+private:
+    float _DT;
+};
+
 class LowPass2pHelper : public LowPassFilter2pFloat {
 public:
     using LowPassFilter2pFloat::LowPassFilter2pFloat;
@@ -67,12 +94,14 @@ public:
 
 
 // create an instance each filter to test
+LowPassConstDtHelper lowpassConstDt;
 LowPassHelper lowpass;
 LowPass2pHelper biquad;
 NotchHelper notch;
 NotchHelper notch2;
 
 enum class filter_type {
+    LowPassConstDT,
     LowPass,
     Biquad,
     Notch,
@@ -97,7 +126,7 @@ void setup()
     const float sample_rate = 1000;
     const float target_freq = 50;
 
-    type = filter_type::Combination;
+    type = filter_type::LowPassConstDT;
 
     // Run 1000 time steps at each frequency
     const uint16_t num_samples = 1000;
@@ -108,6 +137,11 @@ void setup()
     // Print transfer function of filter under test
     hal.console->printf("\n");
     switch (type) {
+    case filter_type::LowPassConstDT:
+        lowpassConstDt.set_cutoff_frequency_override(sample_rate, target_freq);
+        lowpassConstDt.print_transfer_function();
+        break;
+
     case filter_type::LowPass:
         lowpass.set_cutoff_frequency_override(sample_rate, target_freq);
         lowpass.print_transfer_function();
@@ -147,6 +181,7 @@ void setup()
 
 void reset_all()
 {
+    lowpassConstDt.reset(0.0);
     lowpass.reset(0.0);
     biquad.reset(0.0);
     notch.reset();
@@ -156,8 +191,11 @@ void reset_all()
 float apply_to_filter_under_test(float input)
 {
     switch (type) {
+        case filter_type::LowPassConstDT:
+            return lowpassConstDt.apply(input);
+
         case filter_type::LowPass:
-            return lowpass.apply(input);
+            return lowpass.apply_override(input);
 
         case filter_type::Biquad:
             return biquad.apply(input);
@@ -195,6 +233,9 @@ void sweep(uint16_t num_samples,  uint16_t max_freq, float sample_rate)
             hal.console->printf(", %+.9f", output);
         }
         hal.console->printf("\n");
+
+        // Try not to overflow the print buffer
+        hal.scheduler->delay(100);
     }
 }