Blog skyhook logo GPS Alone Doesn't Always Know the Answer

Nov 21, 2017 10:30:00 AM

GPS Alone Doesn't Always Know the Answer

Posted by Nick Knellinger

Read All Posts

At Skyhook we see all types and sizes of companies looking to add location as either a core component or a premium feature. Location is a critical component for a wide range of use cases such as wearable cameras geo-tagging photos and videos, pet tracking, asset tracking, personal fitness wearables, portable gaming consoles, smart vending machines, child or senior citizen personal safety devices, and everything in-between. The one factor that has persisted throughout these engagements, since wearables and IoT began to rise in the past three years, is the varying spectrum of location technology knowledge that companies have.

Many of the conversations I’ve had recently feel like they could have occurred over 10 years ago when smartphones just started to take off. The reason for that is smartphones have leveraged, and still do to this day, hybrid positioning since the iPhone first came out. These location provider systems leverage GPS, Cellular, and Wi-Fi data to provide the most accurate location possible in any given environment. Whether outdoors or indoors, in bad weather, or in a basement or parking garage, the consumer expects that applications know where they are within roughly 20 meters of accuracy (the distance of about 4 car lengths).

While hybrid positioning provides a complete global solution in smartphones, it is not always included when designing new connected devices. This decision occurs for a number of reasons. In some cases the primary environment for a device is outdoors, so location positioning in dense urban and indoor areas is often overlooked. Alternatively, if positioning is considered for those areas, a team may presume that any subsequent indoor or undetermined GPS location can be estimated “well enough” through algorithms, direction of travel or different dead-reckoning techniques.

It sounds extreme, but we also see many cases where the BOM (bill of materials) and revenue projections for a device have already been modeled out for investors or management prior to any sort of prototyping or field testing. It isn't until after the first version of a product has been released or moved into beta testing that the product team realizes the location error is dramatically impacting customer experience. The vast majority of the time though, it’s simply a lack of education.

These nitty-gritty details of location yield (percentage of time a good location can be returned) and accuracy (the actual distance estimated compared to the real device’s location) are often taken for granted. Even the term GPS is constantly used generically as a synonym for location. In fact, GPS (Global Positioning System) is the space-based radio navigation system specifically owned by the United States government and operated by the United States Air Force. It was first made public way back in 1983.

The more general term GNSS - Global Navigation Satellite System - is the standard term for any satellite navigation system that provides autonomous geo-spatial positioning with global coverage. It is true that the quality of GNSS chips and their costs have improved significantly over the past decade. Some of these improvements come as GNSS assistance software continues to evolve and new GNSS satellite constellations from the EU, China, Russia, and Japan have been put into orbit. Even with billions spent to improve the capabilities, “GPS” alone is not always good enough for products that have location as a critical component.  

Beyond core bandwidth and power constraints, here are four things to consider when when designing a product or service with location:

1) Consumer expectations– The average consumer is accustomed to, and expects, smartphone-level global location accuracy from any device they are using. Make sure to remember that when considering how to incorporate location into your device - both on the hardware and software side.

2) Time spent indoors – According to the National Human Activity Pattern Survey, sponsored by the EPA, the average American spends approximately 87% of his or her life indoors. If you want your device to provide location the vast majority of the time, consider adding network location services as a component to your device (Location provided by Wi-Fi or cellular network signals).

3) Switching & blending – In hybrid positioning services, the location provider or service has the ability to ‘switch’ and smooth between different location sources in real time to provide the most accurate location at any given moment. This ‘switch’ occurs by using signal attributes and quality metrics to determine which individual source or combination of sources may be best. One basic example is when a device is in a dense urban environment it can switch between, or even blend, locations from both GPS and Wi-Fi to provide a smooth location track with the least amount of error possible as you traverse through a city.

4) True accuracy requirements – Consider the real accuracy requirements for your use case. While GPS alone may not suit many cases, sometimes sub-20 meter accuracy in general is simply not required. For example, when tracking shipments of perishable or refrigerated goods, the origin and destination may be already known and logged. While other sensors such as temperature and humidity are critical to the success of the shipment, <1.5 KM accuracy to roughly track the asset may be more than enough to estimate progress and approximate arrival times.

Final Takeaway

While GPS alone is enough in some cases, it may not fulfill all location requirements on its own. When designing a new product, it’s important to consider and test real-world conditions to ensure that the device can obtain accurate location even in dense urban and indoor environments or in bad weather. Hybrid positioning is not the only available answer, but it is a proven approach for providing a complete global location solution.

Everything you need to know about IoT and how to make the Internet of Everything a Reality.