10 lessons learned from 50-plus biometric wearable product development cycles

Other Parts Discussed in Post: AFE4410

This is the second in a three-part series on optical heart rate sensors for biometric wearables. The first installment focused on how these sensor systems work and what you can measure with them.

You’re sitting in a planning meeting. The team has just come up with some pretty amazing ideas for the new wearable device with an optical heart rate sensor that you’re adding to your product line. They all turn to you: Who should we partner with for the sensor system? Suddenly you realize that you’re not entirely sure where to begin.

At Valencell, we’ve heard this before — many times, having been involved in more than 50 biometric wearable projects. Valencell uses the AFE4410 in its next-generation Benchmark sensor system and our technology is integrated into more than 20 different biometric wearable devices on the market today. We’ve made some mistakes, learned a lot from those experiences and had some successes along the way. Here are the top 10 lessons we’ve learned throughout the process that you should consider when selecting a biometric sensor system for your next wearable.

1. Get proof that it works before you get started. This might seem pretty obvious, but anyone who has brought a biometric wearable to market will tell you the hardest part of the product development cycle is getting the biometric sensor system to work well enough to meet your requirements. Your starting point should be a working reference design, similar to the one in Figure 1. And don’t just test it on the bench. Put it on at least 20 people and see if it works in your intended use case, compared to a legitimate benchmark device. Another option is to test other companies’ products in the market that are using a particular vendor’s technology. Are their customers pleased with the performance? How does it perform in your tests?

Figure 1: Test a working reference design for biometric wearables

2. Work with a technology provider that knows your market and can support your requirements. Applications for biometric wearables are broad. Make sure your technology provider has the capabilities and experience to deliver on the device you want to build. The right form factor, the right metrics, the right levels of accuracy and the right user experience all play huge roles in a product’s success.

3. Work with a technology provider that has expertise in biometric wearables and optomechanics. Hardware components and optomechanical design are critical factors in accurate biometric wearables. The provider you select should have extensive hardware and optomechanical expertise, and should be experts in evaluating the performance of analog front ends (AFEs) for signal processing, like the TI AFE4410. They should understand how light couples to and from the body to optimize the signals indicating blood flow and minimize environmental noise. A strong team will be sure that the right wavelengths are being used for the chosen device body location and form factor.

4. Advanced signal-extraction methodologies are critically important. Because of the complexities in individual bodies, it’s critical for all technology developers to have a keen understanding of signal extraction algorithms, particularly something called active signal characterization (shown in Figure 2), which identifies and characterizes both motion and environmental noise that can corrupt signals. Isolating these factors, and others, enables the biometric sensor system to report accurate data.

Figure 2: Signal processing involving active signal characterization is critical

5. Advanced biometric measurements are required. Basic heart rate measurement is quickly becoming a minimum requirement for wearables. More advanced metrics will define the future, so take a serious look at what metrics you can get from the system. Does it support continuous heart rate or just take spot measurements periodically? Can it deliver more advanced metrics like beat-to-beat (RR) interval (heart rate variability [HRV]), breathing rate or blood pressure? And, perhaps more importantly, can the vendor prove to you that they can deliver those metrics? Actions speak louder than words.

6. Work with a technology provider who can test your product appropriately. When you partner with vendors for core components of your product, it’s important to know not only how they test their own work, but also how they test their product within your device. If they offer prototype testing services, you’re heading in the right direction. But be sure to ask if they can test your intended use cases, shown in Figure 3. How and where do they conduct that testing? How many people do they test in each round of prototyping?

Figure 3:  Ensure you test the device extensively on the intended use case.

7. Look for proven experience in getting biometric wearables to market. It’s critical to understand what other products on the market are using the tech you’re looking at including in your product. Analyzing consumer reviews of those devices can give you critical insight into how that component will work in your device. How accurate are those products? Are the product reviews positive — at least as they pertain to the heart rate monitoring and biometrics component?

8. Choose a team with multi-disciplinary expertise. You should get an understanding of the skills and expertise of the vendor’s team. Do they specialize in hardware? Optics? Firmware? Algorithms? What are their core competencies?

9. Manufacturing partner experience is highly valuable. Building a biometric wearable is a substantially different project than a typical wearable product, primarily because of the optomechanics involved. Even if you’ve designed and built activity tracking wearables before, don’t underestimate the complexity of adding optical biometrics. You should choose a contract manufacturer that has experience building biometric wearables at scale. They will have faced and solved many unforeseen challenges that other manufacturers have not.

10. Evaluate the intellectual property (IP) landscape. Despite being a relatively new space, there is an extensive patent landscape in biometric wearables and high-profile companies are litigating in this space. The IP landscape for biometric wearables includes areas such as optomechanical designs, data processing, data assessments and visualization of wearable data. Figure 4 highlights the recent trends in wearables patents granted. 

Figure 4:  Wearables patents granted and submitted (photo courtesy of ipwatchdog.com)

I hope this helps you select the right biometric sensor system. The next installment in this series will cover different devices and biometric sensor system use cases.

Additional resources