nan

[maciekish]

Hi,
I'm trying to implement this in my custom location manger on iOS but get_lat_long get_bearing and get_mph all just return nan.

I'm creating the filter like this:

self.kalmanFilter = alloc_filter_velocity2d(self.noise); //Noise is 1.0f

Updating it like this:

update_velocity2d(self.kalmanFilter, location.coordinate.latitude, location.coordinate.longitude, timeIntervalSinceUpdate); //Where timeIntervalSinceUpdate is an NSTimeInterval since the last CLLocation

And reading it like this:

CLLocationDegrees lat, lon;
get_lat_long(self.kalmanFilter, &lat, &lon);
CLLocationDirection bearing = get_bearing(self.kalmanFilter);
CLLocationSpeed speed = get_mph(self.kalmanFilter) * 0.44704f; //Convert mph to meters per second
NSLog(@"%.1f %.1f %1.f %1.f", lat, lon, speed, bearing);

All i get is "nan nan nan nan"

[maciekish]

Turns out this is the culprit:

NSTimeInterval timeIntervalSinceUpdate = [location.timestamp timeIntervalSinceDate:self.lastRawLocation.timestamp];

if (!self.lastRawLocation || timeIntervalSinceUpdate > 0) {
    update_velocity2d(_kalmanFilter, location.coordinate.latitude, location.coordinate.longitude, timeIntervalSinceUpdate);

    self.lastRawLocation = location;
}

If i replace timeIntervalSinceUpdate in update_velocity2d with 1 it start working. What does this parameter do and does it really need to be seconds since last update? What happens if i leave it at 1?

[resqrobert]

@maciekish, any luck with this issue?

[PierrotLL]

Hi, I just used this code (transposed in AS3) and it works very well. Many thanks to the author.

Make sure your values (latitude, longitude, elapsed time) aren't NaN when you update the filter.

I know the guys above will probably never see this answer, but it's more for future viewers.

[GeneralJing]

@PierrotLL can you supply some test examples? I met the same question. below is my code:
// KalmanFilter f = alloc_filter(2, 1);
KalmanFilter f = alloc_filter(7, 4);

/* The train state is a 2d vector containing position and velocity.
    Velocity is measured in position units per timestep units. */

// set_matrix(f.state_transition,
// 1.0, 1.0,
// 0.0, 1.0);
set_matrix(f.state_transition,
1, 0, 0, 0, 1, 0, 0,
0, 1, 0, 0, 0, 1, 0,
0, 0, 1, 0, 0, 0, 1,
0, 0, 0, 1, 0, 0, 0,
0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 1, 0,
0, 0, 0, 0, 0, 0, 1);