#!/usr/bin/env python ''' fit best estimate of magnetometer offsets from ArduCopter flashlog using the algorithm from Bill Premerlani ''' import sys # command line option handling from optparse import OptionParser parser = OptionParser("magfit_flashlog.py [options]") parser.add_option("--verbose", action='store_true', default=False, help="verbose offset output") parser.add_option("--gain", type='float', default=0.01, help="algorithm gain") parser.add_option("--noise", type='float', default=0, help="noise to add") parser.add_option("--max-change", type='float', default=10, help="max step change") parser.add_option("--min-diff", type='float', default=50, help="min mag vector delta") parser.add_option("--history", type='int', default=20, help="how many points to keep") parser.add_option("--repeat", type='int', default=1, help="number of repeats through the data") (opts, args) = parser.parse_args() from pymavlink.rotmat import Vector3 if len(args) < 1: print("Usage: magfit_flashlog.py [options] ") sys.exit(1) def noise(): '''a noise vector''' from random import gauss v = Vector3(gauss(0, 1), gauss(0, 1), gauss(0, 1)) v.normalize() return v * opts.noise def find_offsets(data, ofs): '''find mag offsets by applying Bills "offsets revisited" algorithm on the data This is an implementation of the algorithm from: http://gentlenav.googlecode.com/files/MagnetometerOffsetNullingRevisited.pdf ''' # a limit on the maximum change in each step max_change = opts.max_change # the gain factor for the algorithm gain = opts.gain data2 = [] for d in data: d = d.copy() + noise() d.x = float(int(d.x + 0.5)) d.y = float(int(d.y + 0.5)) d.z = float(int(d.z + 0.5)) data2.append(d) data = data2 history_idx = 0 mag_history = data[0:opts.history] for i in range(opts.history, len(data)): B1 = mag_history[history_idx] + ofs B2 = data[i] + ofs diff = B2 - B1 diff_length = diff.length() if diff_length <= opts.min_diff: # the mag vector hasn't changed enough - we don't get any # information from this history_idx = (history_idx+1) % opts.history continue mag_history[history_idx] = data[i] history_idx = (history_idx+1) % opts.history # equation 6 of Bills paper delta = diff * (gain * (B2.length() - B1.length()) / diff_length) # limit the change from any one reading. This is to prevent # single crazy readings from throwing off the offsets for a long # time delta_length = delta.length() if max_change != 0 and delta_length > max_change: delta *= max_change / delta_length # set the new offsets ofs = ofs - delta if opts.verbose: print(ofs) return ofs def plot_corrected_field(filename, data, offsets): f = open(filename, mode='w') for d in data: corrected = d + offsets f.write("%.1f\n" % corrected.length()) f.close() def magfit(logfile): '''find best magnetometer offset fit to a log file''' print("Processing log %s" % filename) # open the log file flog = open(filename, mode='r') data = [] data_no_motors = [] mag = None offsets = None # now gather all the data for line in flog: if not line.startswith('COMPASS,'): continue line = line.rstrip() line = line.replace(' ', '') a = line.split(',') ofs = Vector3(float(a[4]), float(a[5]), float(a[6])) if offsets is None: initial_offsets = ofs offsets = ofs motor_ofs = Vector3(float(a[7]), float(a[8]), float(a[9])) mag = Vector3(float(a[1]), float(a[2]), float(a[3])) mag = mag - offsets data.append(mag) data_no_motors.append(mag - motor_ofs) print("Extracted %u data points" % len(data)) print("Current offsets: %s" % initial_offsets) # run the fitting algorithm ofs = initial_offsets for r in range(opts.repeat): ofs = find_offsets(data, ofs) plot_corrected_field('plot.dat', data, ofs) plot_corrected_field('initial.dat', data, initial_offsets) plot_corrected_field('zero.dat', data, Vector3(0,0,0)) plot_corrected_field('hand.dat', data, Vector3(-25,-8,-2)) plot_corrected_field('zero-no-motors.dat', data_no_motors, Vector3(0,0,0)) print('Loop %u offsets %s' % (r, ofs)) sys.stdout.flush() print("New offsets: %s" % ofs) total = 0.0 for filename in args: magfit(filename)