What every IoT engineer needs to know about navigation: GPS, dead reckoning, cellular, Wi-Fi, and BLE
David Scheltema
Posted on May 15, 2019
Personal navigation has never been easier. With numerous options from GPS and cellular to Wi-Fi and Bluetooth, it can be challenging to select the right one for your product be it micromobility scooter or a fleet of logging trucks. To understand which solution is right for you, it’s helpful to have a high-level perspective on the most popular and widespread geolocation technologies.
Most geolocation solutions should start with GPS since it’s a system designed from the start to answer the question, where am I? With close to ~4-meter accuracy, it’s a no brainer to start any navigation solution here, unless of course, your use-case is underground.
Unsurprisingly, no single navigation solution is right for every product use-case or fleet deployment. What works for a micromobility scooter — a combination of GPS, cellular, and Bluetooth — would be an over-engineered solution to keep track of your keys. For that, something like Bluetooth beacons are more appropriate.
This guide will provide you with information about commonly radio technology used for navigation. After reading, you should be familiar with the different available technologies as well as a high-level awareness of the significant trade-offs between the options. In many cases, the best solution is a robust combination of technologies.
Navigation options at a glance
Determining the right option for your geolocation needs is a balancing act between the cost of the silicon and the accuracy of the location data it provides. Battery life is not covered in this overview, however, it’s good to keep in mind that the number of times you send location data from your device to the cloud will affect your battery performance. Additionally, specific radio power performance will vary by vendor and part.
GPS — The Global Positioning System (GPS) is generally the most accurate, and it can provide device location continuously. This option is moderate to high priced depending on features such as onboard RTKs. It is widely used used as the basis for any geolocation solution.
Cellular — Cellular geolocation uses known cell tower location to approximate where you are. Of course, cellular requires an active service plan and proximity to cellular towers. Cellular is predominately used as a supplement to GPS, as it lacks the accuracy to be an independent solution.
Wi-Fi — Wi-Fi is not advisable for use-cases that move a lot, but is commonly found in smartphones and laptops as a way to supplement location data or give approximate location inside of buildings using 3rd party databases of Wi-Fi access point locations.
Bluetooth 5/BLE — Great for small tracking devices such as key fobs and assets in a contained, small location. Typically deployments use a mobile phone in conjunction with Bluetooth beacons that broadcast location information.
Getting started: Weighing cost vs accuracy with GPS
Unless your use-case is subterranean, your navigation solution should begin with GPS. Thanks to smartphones, the cost of GPS modules is lower than it’s ever been, but that means there are a lot of module options to consider for your design. A helpful way to cut through the options is by balancing between unit cost and location accuracy.
As a general rule of thumb as GPS accuracy increases, so too does the cost. Inexpensive GPS modules can be purchased for under $40 and typically support UART or I2C depending on the module. More expensive, featureful GPS modules can run upwards of $200, but offer advanced capabilities like onboard RTKs. However, there’s also a bit of a middle ground. To supplement the accuracy of GPS, it’s not uncommon to see silicon vendors and product designers add additional hardware or using software to help gain higher accuracy.
Using cellular plus geofencing for robust geolocation
Cellular is typically a medium for voice and data, but you can also use it to determine approximate location. Note that a cellular-only navigation system will only provide you with your location relative to the towers your cellular device is communicating with. For more in-depth information on cellular, please see our cellular guide in the Particle docs.
Every cell tower has an identifier, and this can be looked up using Google’s Geolocation API to find where it is on the globe. This process is fast and provides a general location, usually within 4,000 meters or a couple of miles. Of course, you need a cellular signal for this to happen.
Pairing cellular connectivity with GPS-based geofencing provides a solid solution for IoT deployments that may go in and out of hard to reach places such as lumber trucks across the state of South Carolina. Particle’s customer Worthwhile combines these two technologies to provide a more robust navigation solution. Using a Particle cellular solution their open-source Geo-Bound software caches up to 2,000 GPS coordinates locally in their asset trackers memory and stores them until back in cellular range. While cellular here is not used for the navigation per se, the combinational approach of cellular plus GPS highlights a robust geolocation solution.
Improving GPS accuracy with IMU-based dead reckoning
In some situations, GPS alone cannot provide a full navigation solution. Take a fleet of scooters for example. In big cities they don’t always have a clear line of sight to the sky for GPS to work efficiently — gaining a lock on satellites takes time and consumes power, both challenging in urban environments.
By augmenting a GPS module with additional geolocational data and using techniques like dead reckoning — a navigation technique that uses information about your past position, speed, duration, and heading to calculate where you currently are — a fleet for scooters can still be found even if they can’t see the sky.
Developers use dead reckoning to supplement GPS data and help with error correction. To cost effective achieve dead reckoning, it’s not uncommon to see prototypers add an inertial measurement unit (IMU), a specialized sensor that’s packaged in a single integrated circuit, to collect information about the heading and speed information as a supplement the available GPS coordinates.
Dead reckoning works best with micromobility, specifically when tracking mobile assets around an urban area, with inexpensive non-GPS navigation devices, and it’s very accessible for DIY robotics projects.
While alone, an IMU-based dead reckoning solution isn’t recommended for high-value asset tracking, supplementing a GPS with an IMU can result in a more robust geolocation solution.
Dead reckoning in silicon: high-end GPS modules with built-in real-time kinematic engines
Land surveying and autopilot systems use GPS that include real-time kinematic engines (RTK) that perform dead reckoning inside the GPS module and boast centimeter accuracy. No extra ICs required. One example of such a module is the ublox NEO-M8P-2, which SparkFun offers in a sub-$200 breakout board.
The only problem is GPS modules with RTKs is the cost. Single quantity pricing for GPS modules with integrated RTK run upwards of $200 per unit, making them too expensive for most use-cases. Unless you’re building a centimeter accurate navigation solution like surveying a building site or autonomously assisting an autopilot system, using a GPS with RTK is over-engineering your solution.
As you can probably guess, due to the higher per unit cost, GPS modules with on-die RTKs are almost non-existent in micromobility markets, but more common in industries small measurements matter. For those undeterred by the unit price, Sparkfun electronics has an excellent primer on GPS with RTK, which also covers more advanced navigation techniques such as error correction using Radio Technical Commission for Maritime.
Wi-Fi provides supplemental navigation
Smartphones are a fantastic example of Wi-Fi supplementing other radios for geolocation. You’ve probably experienced your smartphone reminding you to turn on WI-Fi to help improve your location accuracy.
Wi-Fi assisted navigation works by using special 3rd-party API service such as Google’s Geolocation API that provides access to an enormous database of Wi-Fi access point locations. Given the scale of their enterprise and given enough data points of Wi-FI APs, the information is helpful for indoor navigation.
Problems with using Wi-Fi for navigation
One enormous challenge with using Wi-Fi as a navigational aid is that as soon as you move out of range of a Wi-Fi access point, your access to the Wi-Fi location database is lost. Smartphones do a fantastic job of obscuring this limitation since they have cellular, GPS, and Wi-Fi that work seamlessly together and require near-zero user configuration.
Much like cellular, it’s not uncommon to see early prototype projects attempting to combine Wi-Fi with an IMU to approximate a navigation system. These systems are architecturally similar to cellular plus IMU approach, but with one major drawback: limited range bound by Wi-Fi signal coverage. Without a Wi-Fi signal, there is no way for the device to connect to a cloud for dashboards, fleet management, or even access the geolocation APIs.
Don’t combine cellular and Wi-Fi to navigate unless you absolutely must
Some might think to add cellular and have a Wi-Fi and cellular solution, however, unless you’re making a smartphone, this is not an advisable approach to navigation. Adding a second radio interface such as cellular means you effectively double your power draw, increased bill of materials cost, your software integration time and rigor, and your antenna design just got a lot more complicated. Instead, a navigation solution that combines GPS with Wi-Fi is a much more appropriate path.
Bluetooth navigation handles the small stuff like finding your keys
Geolocation isn’t just about getting you from your home to office. In some use-cases, it’s about finding your misplaced keys. With a limited range of 10 meters, Bluetooth is great for helping you track down things you know are nearby but needs some help finding.
Increasingly Bluetooth 5 based devices featuring ICs like the Nordic nRF82540 found in all Particle 3rd Gen devices are used for key-finding solutions. Of particular interest to designers is the improved and descriptively named Bluetooth broadcast and beacon modes. These allow for information-rich interactions between non-paired Bluetooth devices.
For example, imagine a rideshare bike that could broadcast its battery state directly to the customer. No need to send data to the cloud, simply use a beacon mode to advertise to the rideshare app the charge level. For more on the use of Bluetooth in navigation, please see the Bluetooth SIG documentation here.
Navigating your way to the right solution for your use-case
Ultimately, the best way to understand each of these technologies is with hands-on experience. As you’ve learned, the most reliable geolocation and navigation solutions use a robust combination of radio technologies. For your own prototyping journey using a Gen 3 Particle device with an Adafruit Ultimate GPS FeatherWing is a great place to start. From there, it’s very easy to transition to an enterprise-grade module like our B Series SoM for scale.
For more on how Particle solutions can help with your geolocation and navigation products see our product suite overview or contact us and talk with an expert.
The post What every IoT engineer needs to know about navigation: GPS, dead reckoning, cellular, Wi-Fi, and BLE appeared first on Particle Blog.
Posted on May 15, 2019
Join Our Newsletter. No Spam, Only the good stuff.
Sign up to receive the latest update from our blog.
Related
May 15, 2019