The availability of low-cost GPS jamming and spoofing technologies renders GPS-only solutions for location and navigation an increasingly dangerous choice for the dismounted soldier in a battlefield environment. Therefore, it is essential that a self-contained, non-infrastructure-based location technology be developed to maintain the optimal war fighting capability of the modern US dismounted soldier. However, inertial solutions are very difficult to implement properly even without the unique challenges particularly presented by human motion dynamics. For example, on a battlefield, soldiers tend to dodge, dive, duck, and run, thus creating motion measurement challenges that would cause tracking errors even in existing vehicular inertial navigation units, which are generally too large, heavy and consume too much power for a soldier-based application.


For decades, GPS has been the go-to tracking and navigation solution for soldiers on the battlefield. Unfortunately, most enemies now have access to very cheap and widely available technologies to jam GPS location signals. But even more dangerous, not only can the GPS signals be jammed, they can also be spoofed as well. In the worst cases, the enemy could move soldiers into dangerously exposed positions, or could even have the soldiers inadvertently calling supporting fire onto their very own positions. Needless to say, this is a major problem that needs a solution.

MEMS Navigation

So, given such application requirements, the answer lies in some form of inertial measurement solution. Many Inertial Measurement Units (IMU) already exist on the market, but none can yet sufficiently solve the problem of inertial pedestrian tracking. They are an attractive part of the ultimate solution because they only rely on measuring the forces generated by the user’s movements but operate independently of infrastructure technologies such as WiFi, BLE beacons, or other RF or camera-based anchoring technologies — they can’t be jammed, spoofed, or led astray by a bad actor manipulating the operational environment.

However, no IMU exists presently that can adequately allow for a true integration of a user’s accelerations to track location, while at the same time not be thrown off course by the nature of a soldier’s movements on the battlefield. By increasing sensitivity of the inertial sensors to accurately measure small accelerations produced by human locomotion, the susceptibility to the other artifacts of human locomotion creates significant errors on the opposing side of the sensors’ measurement spectra. That is, saturation and other high-force and rapid-movement induced errors create even more significant errors, which makes this kind of tracking so difficult to do.

Ideally, MEMS inertial sensors would be perfect for a dismounted soldier application because they’re small, lightweight, low power and low cost. In other words, they have all the most desired characteristics of low SWaP-C (Size, Weight and Power-Cost). Every soldier could easily carry one or more — on a wristwatch, in a backpack, on a helmet or mounted to a weapon. However, their performance in terms of sensitivity, noise, bias drift, and other related error modes, make them perform very poorly for inertial navigation. Within a few minutes, positioning errors of up to hundreds or even thousands of meters could accumulate, thus rendering their inertial navigation efficacy unusable. However, they could be reasonably applied to location tracking through dead reckoning.

Figure 2. PNI Sensor TRAX2 AHRS positive & negative pitch and roll definition. (Image courtesy of PNI Sensor)

In its simplest form dead reckoning tracks a user’s position by measuring how far the user has traveled in a series of directions. If the IMU can provide a relatively accurate direction with respect to North (magnetic compassing) and then measure how far the user has traveled (step counting), a fairly accurate position can be constructed. There are many challenges in producing an accurate and reliable MEMS dead reckoning system, but at least the application does not push the performance boundaries of the sensors themselves.

A MEMS gyroscope is a disc that is micro-machined out of silicon that is excited to vibrate at a rate that is significantly higher than the bandwidth of the motion to be measured. As the disc is being vibrated along one axis, if there is a rotation along its sensitive axis, a Coriolis force will be produced that sends the vibrating force into an orthogonal axis which is measured and is proportional to the angular velocity of the rotation. The amplitude of the signal corresponding to that rotational rate is quite small, however, and it is not a direct measurement of displacement but of rate of displacement over time, which then needs to be integrated into an angular displacement. However, it only requires a single integration, not the double integration needed to determine linear displacement using an accelerometer.

The gyro is used solely to measure relative rotation with respect to an arbitrary starting point. It is immune (or nearly so) to linear acceleration, so it helps provide very precise rotational measurements, irrespective of shock and vibration. Because of its measurement characteristics, it provides accurate angular displacement, with very low latency and overshoot.

From Point A to Point B

The IMU can accurately determine the direction of travel and thereby track the path taken from the time a soldier has left point A. An odometry function is needed to measure how far they’ve traveled in any given direction as they progress towards point B. An odometer for a car is relatively straightforward. There is a sensor that counts the number of partial and complete rotations of the car’s wheels and that number is multiplied by the circumference of the wheels to derive a fairly accurate distance travelled. Modern odometers can be accurate to within 0.1% of the actual distance the car was driven. A similar function for people would simply be a pedometer, or step counter. Step counters have improved dramatically over the last five years due to activity devices, smart watches, and cell phones. It turns out that people walk with fairly consistent stride lengths, so multiplying the number of steps someone takes by their stride length can yield distance traveled accuracies of 1% or better.

Figure 3. PNI Sensor TRAX2 AHRS & digital compass module. (Image courtesy of PNI Sensor)

The complexity of this approach is to determine the different stride lengths for walking, jogging, running and even crawling. There are several available approaches for determining the stride length fairly well — for instance, counting steps in open sky while there is GPS coverage and then dividing the distance traveled by the number of steps taken between the two endpoints. In any case, it is a solvable problem and just depends upon what an acceptable procedure is for the specific use cases.

Navigating in Combat

The current generation of best-of-breed MEMS gyros can now track soldiers’ dynamic motion very well for up to 30 minutes without significant bias or scale errors. The only thing that the system we developed can’t do with them just yet, is to lock on to North using gyro-compassing techniques, since their 1/f noise is still about a factor of 50 too high. MEMS are really good for tracking highly dynamic motion in a short period of time like playing with your Wii controller. But for applications where you need low noise and high sensitivity, for example measuring the Earth’s rate of rotation of 15° per hour, which is a very small signal, it’s not quite there yet.

Finding the Coordinates

Absolute geographical location can be pinpointed by the intersection of latitude and longitude.


The IMU’s accelerometer can directly measure the direction of the gravity acceleration vector, which means that it can accurately determine the direction of down (towards the center of Earth). The gyro can accurately measure the X, Y, and Z components of the angular velocity of the Earth’s rotation at its given location. Then, the latitude can easily be determined by using these two measurements.


Longitude can be calculated by measuring the deviation angle between the measured geographic north and magnetic north. The two are aligned along different axes and so, there’s always a deviation in terms of direction between them at your longitude. Since both measurements are made in three dimensions (the axis of the Earth’s rotation and the flux lines generated from the axis of the Earth’s magnetic dipole), you could convert the difference in three-dimensional space between those two in order to locate yourself longitudinally on the surface of a sphere, especially if the latitude has already been determined by the previous measurements.

An IMU with a high-grade gyro can locate true north via gyro-compassing (or North seeking), but it takes a magnetic compass to locate magnetic north. However, even though there’s a lot to be said for a magnetic sensor giving you a direct reading to north, there are serious problems with how fallible the measurement can be. Aside from calibrating for hard and soft iron interference on the user’s body, external magnetic distortions are often present that are difficult to account for. For example, if a soldier walks by building wall with iron rebar, it can create large errors in the magnetic measurement. And in an urban warfare environment, there are many additional sources of interference including vehicles, buildings, weapons, ordinance, and many other structures a soldier is likely to encounter.

PNI Sensor’s Solution

PNI Sensor (Santa Rosa, CA) is in the development stage of a tracking device that combines the best elements of the newest generation MEMS devices with an electronic compass that uses the most advanced magnetic anomaly detection and rejection algorithms on the market. Based upon our TRAX2 attitude and heading reference system (AHRS), it employs a unique Kalman algorithm that intelligently fuses PNI’s reference magnetic sensors with gyros and accelerometers. In addition, PNI Sensor has developed some of the most advanced pedometry functionality for use in its tracking device for very high dead reckoning tracking performance for mission critical applications.

This article was written by George Hsu, CTO, PNI Sensors (Santa Rosa, CA). For more information, contact Mr. Hsu at This email address is being protected from spambots. You need JavaScript enabled to view it. or visit here .