Accuracy down to a centimeter or two for satellite tracking opens the door to new applications including autonomous driving, professional drone navigation and wireless industrial robotics control.
The current level of accuracy for GNSS (Global Navigation Satellite Systems), of between one and five meters, is fair enough for satellite navigation systems in vehicles, Google maps and recreational tracking as in sports watches, although even some of these would benefit from an improvement. But for collision avoidance for drones and vehicles, and even more so for high precision process control, much greater improvement down to at least one or two centimeters accuracy is required.
It is true that the latest generation GNSS systems, such as the European Union’s Galileo, boast improved accuracy, but this is still in the order of a meter, which is enough to satisfy, say, users of sports watches for most scenarios, but not sufficient for those other cases. However, an enhancement availability to most GNSS systems promises to deliver the desired accuracy, or close to it. This is ‘real-time kinematics’ (RTK), the technology now commercially available in the A from Verizon via its ThingSpace’s IoT platform under the banner of Hyper Precise Location (HPL).
It may not be quite hyper, but it is a two-magnitude improvement over GNSS as it stands, demonstrating around 1cm accuracy horizontally and 2cm in the vertical dimension. Verizon is pitching it particularly at the automotive and industrial IoT sectors.
It will play a pivotal role in the operator’s collaboration with Honda, announced early this month, to evaluate how edge-based 5G can enable reliable connectivity between road infrastructure, vehicles, and pedestrians. The two are working at the University of Michigan’s Mcity, a testbed for connected and autonomous vehicles, on the project, with a remit to “reduce collisions and save lives”. Verizon is providing the 5G connectivity and mobile edge computing (MEC), while Honda’s contribution is its SAFE SWARM the automotive safety platform, based on cellular vehicle-to-everything (C-V2X) to share data about location and speed of traffic, coupled with vehicle sensor data.
RTK is effectively an enhancement layer to GNSS, using the phase of the satellite signal’s carrier wave to derive a more precise determination of its transit time and therefore distance traversed. With conventional GNSS positioning, signals are received from several GNSS receivers and location is calculated by first measuring the differential in signal transit times from each one. Then trigonometry is applied to combine these measurements with the known locations of those satellites in their orbits at that point in time to determine the device’s position.
To calculate those signal delays from the various satellites, the receiver aligns a pseudo-random binary sequence contained in each signal to an equivalent internally generated sequence, the aim being to align them. By progressively incrementing the receiver’s sequence, the two sequences are eventually aligned and the time taken for this process to converge is why there is some delay booting up a satellite navigation system.
There are errors inherent in this process that can be reduced but not eliminated. One lies in the ability of the receiver’s electronics to process signals from the satellite accurately, but there are also error sources associated with the transmission, such as non-mitigated ionospheric and tropospheric delays, as well as ephemeris errors resulting from uncertainty over the exact position of each satellite.
RTK operates along the same trigonometric principles, but cuts down on the errors associated with signal transmission by appointing a fixed base station to transmit correction data locally to the roaming device. While the mathematical details of implementation are quite complex, the underlying principles are relatively straightforward.
The key points are that the location of the fixed based station is known, while the biggest errors in GNSS signal processing are also constant. Therefore those errors can be cancelled out by applying differential processing to signals received by the roaming device and the fixed station, if they are sufficiently close together.
This proximity of device and base station also allows use of phase which confers greater accuracy but again requires differential measurements to determine. The main problem is that while it is easy to determine the phase of a wave, that is the position within its oscillating cycle, it is not possible to be sure which cycle of the wave is being received at the time. So the measurement could be wrong by some integer multiple of the wavelength, known as the carrier phase ambiguity. This can be resolved through comparison of phase with the fixed base station whose distance from the moving device is known accurately enough.
Interestingly, computation of the base station’s fixed position is itself a constraint over even greater accuracy. The relative position of the device to the base station can be determined to within around 2mm, which implies there is potential for further improvements. This also implies that some applications that only need relative rather than absolute positions could benefit from this even greater accuracy now, and that includes mapping surveys.
One caveat is that RTK relies on some form of internet access on the device. This can be a constraint for some use cases such as vehicle satellite navigation, whose enduring popularity rests partly on providing ubiquitous coverage even when a vehicle is beyond the range of a cellular signal in remote areas.
However, in the automotive case, fall back to traditional GNSS accuracy in relative rare situations when out of range of a base station is probably not too great a handicap.
Thomas Fox, SVP of industrial IoT and automotive for Verizon Business, was quite bullish. “HPL stands to boost or enable next-gen technologies across industries, from intelligent-driving to drone delivery to highly automated operations within construction, agriculture, and much more,” said Fox. “HPL’s fast expanding coverage area, API friendliness, privacy protection, and use of open-delivery standards make it ideal for developers and customers demanding precision and flexibility.”