Tern AI develops low-cost navigation alternative to reduce reliance on GPS

Many important systems in the contemporary world rely on GPS, such as aviation and road networks, emergency and accident response, etc., but GPS signals are susceptible to interference, destruction and deception. According to foreign media reports, the startup Tern AI proposed to create an independent derivative positioning system (IDPS) that can identify the location of vehicles or personnel without relying on satellite signals. This means that there is no interference, construction or blind spot obstacles that can affect precise positioning, and it is expected to provide a GPS alternative.

(Image source: Tern AI)

“We can do the job of GPS without having to call up to space and ask, ‘Where am I?’” said Shaun Moore, CEO and co-founder of Tern AI.

The current system works by having a GPS receiver in a car or phone receive signals from satellites orbiting the Earth, which contain information about the satellites and the precise time the signals were sent. The GPS receiver then uses the time it takes to transmit each signal to calculate its distance to each satellite. By triangulating the distances, the receiver can determine its exact location, which is then displayed on apps like Google Maps or Waze. This approach creates a globally standardized positioning technology, but it is also susceptible to local interference. Therefore, a more secure solution is needed for a car's GPS system.

Tern AI's IDPS is powered by an artificial intelligence (AI) model that ingests vehicle and phone sensor data in real time, as well as third-party map data, to provide consistent and precise locations. The model can be downloaded as a software package to the infotainment system of compatible vehicles. If the vehicle is not compatible, the system can be downloaded to a smartphone. Tern's model can utilize phone sensor data to help determine location, but in some cases, the smartphone needs to pull car sensor data from the vehicle's onboard diagnostic port.

Moore said the goal is to have leading mapping products like Apple Maps, Google Maps or Waze get location information from Tern's AI model instead of from GPS.

The biggest advantage of Tern AI is that it can use artificial intelligence to interpret data from sensors already on smartphones and vehicles, such as speed and motion sensors. Moore pointed out that Tern does not rely on computer vision and cameras to help it determine positioning.

Moore noted that while GPS has improved, the fundamental nature of the technology has not changed, nor has it evolved enough to protect GPS from attacks. Other alternatives to GPS satellites, such as low-Earth orbit satellites and ground-based beacons, while impressive, are very expensive to install and maintain. "GPS technology has not changed significantly in 50 years, and other solutions to address or mitigate risks have only achieved minor improvements, not real innovation," he said. "One of our founding principles is 'No Signal,' which means we don't have to look for help from the outside to determine location. Now, just add intelligence and the device you have can do that."

Tern validates its technology primarily through internal testing based on the U.S. Department of Transportation’s framework. Moore said Tern’s IDPS can locate vehicles within 4 meters of accuracy, which meets the current U.S. Department of Transportation guidelines for emergency services.

The startup is set to begin testing with its launch partner (which it says is a giant in logistics, transport and delivery) in July, and is in talks with emergency services and autonomous last-mile delivery companies. Tern plans to launch commercially in September.