过滤指南针读数

时间:2011-03-18 03:12:10

标签: iphone filter mkmapview compass-geolocation moving-average

我正在使用罗盘标题来旋转MKMapView。旋转有点生涩,所以我试图像iPhone上的谷歌地图那样过滤它(或者似乎做了一些诡计)。

我正在尝试使用移动平均公式来过滤iphone指南针的读数,但它在359和0之间的交叉处失败,因为它从35x开始向后平均到0并导致地图在接近时向后旋转从西边向北。

任何想法最好的方法是过滤这些数据,使其从359回到零,并维持滚动平均值。

代码在这里:

- (void)locationManager:(CLLocationManager *)manager didUpdateHeading:(CLHeading *)newHeading {
static float xd=0;
static float k = 0.22;

// Moving average formula
xd = k * xd + (1.0 - k) * newHeading.magneticHeading;

NSLog(@"%0.2f : %0.2f", newHeading.magneticHeading, xd);    
[map setTransform:CGAffineTransformMakeRotation((-1 * xd * M_PI) /180)];}

感谢您的帮助

4 个答案:

答案 0 :(得分:4)

- (void)locationManager:(CLLocationManager *)manager didUpdateHeading:(CLHeading *)heading
{
    int newHeading;
    int queuecount;

    newHeading = heading.trueHeading;
    [queue addObject:[NSNumber numberWithInt:newHeading]];

    if([queue count] > 10) [queue removeObjectAtIndex:0];
    queuecount = [queue count];

    NSEnumerator *e = [queue objectEnumerator];
    NSNumber *sum;
    int oldd = 0 , newd, average =0;
    BOOL firstLoop = YES;
    while ((sum = [e nextObject])) 
    {
        newd = [sum intValue];
        if(firstLoop) {oldd = newd;firstLoop=NO;}

        if((newd +180) < oldd)
        {
            newd +=360; oldd = newd;
            average = average + newd;
            continue;
        }
        if((newd - 180) > oldd) 
        {
            newd -=360;oldd = newd;
            average = average + newd;
            continue;
        }

        average = average + newd;
        oldd = newd;

    }
    average = (average / queuecount) % 360;

    [map setTransform:CGAffineTransformMakeRotation((-1 * average * M_PI) /180)];

}

答案 1 :(得分:2)

如果之前的移动平均线和新标题的差异超过180度,则将360添加到较小的值。然后在存储新移动平均线时由360修改。所以(没有精确的数学):

HDG   MA
350   350
355   353
  0   356  (because 353 - 0 > 180 so adjusted HDG is 360)
  5   359  (likewise)
 10     2  (likewise, then 362 is new MA, mod 360 to normalize)
350   356  (because 2 - 350 < -180 so adjusted MA is 362)

我希望这是有效的,并且比Averaging angles中描述的三角方法更有效(感谢Mark Ransom提到它)。

答案 2 :(得分:0)

我只想添加我的d0n发布的代码版本,并进行一些调整。

我的代码是Arduino的C代码所以我没有使用库(例如我得到没有abs函数的绝对值)...

这个帖子对我很有帮助。谢谢D0n和John ......

int Robot::getHeadingAverage() {

  // setup
  int lastReading = 0;
  int newReading = 0;
  int totalHeadings = 0;
  int avgHeading = 0;

  // loop through all the readings
  for(int i=0; i<totalHeadingReadings; i++) {
    // get the reading
    newReading = headings[i];

    // make sure we have an old reading
    if(i==0) {
      lastReading = newReading;
    }

    if((newReading + 180) < lastReading) {
      newReading = newReading + 360;
    }
    if((newReading - 180) > lastReading) {
      newReading = newReading - 360;
    }

    lastReading = newReading;
    totalHeadings = totalHeadings + newReading;

  }

  // get the average and make sure we do not end up over 360
  avgHeading = (totalHeadings / totalHeadingReadings) % 360;

  // check for negative value
  if(avgHeading < 0) {
    avgHeading = -1 * avgHeading; 
  }

  return avgHeading; 
}

答案 3 :(得分:0)

最后,这对我有用。

libname out 'C:\users\den\Desktop\' ;
proc copy inlib=work outlib=out ;
  select export_project;
run;