Key takeaway: PawPulse originally prototyped with PPG optical sensors - the same technology used in every human fitness tracker and the only sensor type used by competing pet wearables. After months of testing across 15 dog breeds and 6 cat breeds, PPG returned usable heart rate data less than 30% of the time on medium and long-coated breeds. We removed it entirely and rebuilt the platform around millimeter-wave radar, which reads vitals through fur at over 95% reliability regardless of coat type.
We Tried the Easy Path. It Didn't Work.
Every smart collar startup starts with the same assumption: use PPG. It's cheap. It's proven. The supply chain is mature. Human fitness trackers have used photoplethysmography for over a decade, and the component costs are measured in cents, not dollars. If you're building a pet wearable on a timeline and a budget, PPG is the obvious choice.
We made the same assumption. Our first three prototypes had green-light PPG sensors pressed against the underside of the collar, positioned to sit against the dog's neck. On paper, it made perfect sense. In practice, it was a disaster.
Not immediately, though. That's the insidious part. On a short-haired Boxer in a climate-controlled lab, the PPG data looked great. Clean waveforms, stable heart rate readings, publishable signal quality. We high-fived. We moved forward. We started planning production.
Then we tested on a Golden Retriever. Then a Husky. Then a Maine Coon.
And the data fell apart.
The Five Ways PPG Failed Our Dogs
We didn't remove the optical sensor on a whim. We removed it after systematically documenting every failure mode across months of real-world testing. Here's what we found.
1. Fur Blocks the Light (The Big One)
PPG works by shining an LED into tissue and measuring the light that bounces back. The signal relies on photons reaching blood vessels and returning to the photodetector. Fur - even a single layer of it - reflects and scatters that light before it gets anywhere near the skin.
Research published in Scientific Reports (Nature) on PPG signal quality confirmed that even variations in contact pressure between skin and sensor distort the signal morphology. On a furred animal, the contact isn't just variable - it's absent. The sensor is measuring light bouncing off keratin, not blood volume changes.
In our testing, PPG accuracy on short-coated breeds in still positions was roughly 70%. On medium coats, it dropped to under 40%. On long or double-coated breeds (Huskies, Samoyeds, Collies), it was below 15%. On cats - with their denser fur and lower tolerance for tight collars - it was essentially unusable.
2. Motion Killed Everything
Even on the rare occasions we got a clean signal through fur, movement destroyed it. A dog walking, scratching, shaking, or shifting its collar position introduced motion artifacts that overwhelmed the PPG waveform. Published research confirms that motion artifacts are the primary reliability challenge for PPG in wearable devices, with studies documenting that even human wrist-worn PPG sensors struggle during physical activity.
Dogs aren't sitting still at a desk all day. They're running, rolling, digging, and doing whatever dogs do. Research published in Sensors (MDPI) confirmed that PPG accuracy varies significantly with both contact pressure and physical activity level, with higher activity intensities producing larger measurement errors. A sensor that only works when the animal is motionless and the collar is perfectly positioned isn't a health monitor - it's a lab instrument pretending to be a consumer product.
3. Ambient Light Interfered
PPG photodetectors can't distinguish between light from the LED and light from the environment. Direct sunlight, bright indoor lighting, and even reflections off light-colored surfaces introduce baseline noise that degrades signal quality. For a collar worn outdoors half the day, this was a persistent problem.
4. Moisture and Debris Caused Dropout
Mud, water, saliva, and general outdoor grime on the sensor window created optical interference. A dog that swam, rolled in dirt, or drooled on its collar (so, every dog) would produce corrupted readings until the sensor was manually cleaned. That's not a viable workflow for a product owners should be able to forget about once it's on.
5. Contact Pressure Was Impossible to Maintain
PPG accuracy is directly dependent on consistent contact pressure between the sensor and the skin. A study published in PMC demonstrated that PPG measurement accuracy varies significantly with contact pressure, and that optimal pressure differs between individuals. On a dog's neck - where the collar shifts with every head turn, scratch, and play session - maintaining consistent contact pressure is physically impossible.
The Decision That Changed Everything
After four months of testing, we had enough data to see the pattern clearly. PPG wasn't a sensor with some edge cases to work around. It was fundamentally the wrong technology for furred animals.
We had two options:
Option A: Ship with PPG anyway. Market it as "heart rate monitoring," include fine print about coat compatibility, and accept that the feature would only work reliably on 20-30% of dogs. This is, frankly, what the rest of the market does.
Option B: Scrap the sensor, absorb the sunk cost, delay the timeline, and find a technology that actually works on all dogs and cats regardless of coat.
We chose Option B. It was expensive. It set us back months. And it was the best engineering decision we've ever made.
Why Radar Was the Answer
Millimeter-wave radar doesn't use light. It transmits electromagnetic waves that pass through non-metallic materials - fur, fabric, plastic, mud, water - with minimal signal loss. Instead of trying to measure blood volume optically, it measures the micro-mechanical movements of the chest wall: the subtle displacement from each heartbeat (as small as 100 micrometers) and the rhythmic expansion of breathing.
This sidesteps every failure mode that killed PPG:
| PPG Failure Mode | Radar Response |
|---|---|
| Fur blocks light | Electromagnetic waves pass through fur at this frequency |
| Motion artifacts | Signal processing isolates chest displacement from gross body movement |
| Ambient light interference | Radar operates independently of optical conditions |
| Moisture and debris | Electromagnetic waves unaffected by surface contaminants |
| Contact pressure dependency | Non-contact sensing - no skin contact required |
The first time we tested a radar prototype on the same Husky that had produced nothing but PPG noise for weeks, we got a clean heart rate signal in under 10 seconds. That was the moment we knew.
The Trade-Offs We Accepted
We'd be dishonest if we said switching to radar was purely upside. There were real trade-offs, and we think they're worth acknowledging.
Higher component cost. A radar module costs significantly more than a PPG LED and photodetector. This increases our BOM and, eventually, the retail price. We decided that a sensor that works is worth more than a cheap sensor that doesn't.
More complex signal processing. Extracting heart rate from radar return signals requires more sophisticated DSP than PPG's relatively simple peak detection. We invested heavily in firmware and algorithm development. The upside: our signal processing now extracts heart rate, respiratory rate, and HRV simultaneously from the same radar data, which PPG couldn't do even when it worked.
Smaller vendor ecosystem. The supply chain for millimeter-wave radar in wearable form factors is younger and less mature than PPG. We're working with fewer suppliers, which requires more careful component planning. This is a supply chain risk we actively manage.
Higher power consumption per measurement. Individual radar measurements draw more power than PPG. We compensated with smarter duty cycling and power management architecture, targeting a 30+ day battery life despite the more power-hungry sensor.
Every trade-off pointed the same direction: harder engineering, better results. We'd rather solve hard problems than ship easy compromises.
What This Means for Your Pet
If you're evaluating smart collars, the sensor technology isn't a spec sheet footnote. It's the foundation everything else sits on. Prism Insights AI can't build an accurate baseline if the sensor drops out every time your dog moves. Pulse Therapy can't auto-activate on rising anxiety if the heart rate data is unreliable. The Vet Dashboard can't show your veterinarian a meaningful trend if half the data points are missing.
PawPulse's radar-based sensing platform is why every other feature in the Lucero collar actually works. The sensor isn't the product. But without the right sensor, there is no product.
The Industry Will Follow. Eventually.
Every pet wearable currently using PPG for vitals monitoring faces the same physics problem we documented. Fur blocks light. Contact pressure is impossible to maintain on a dog's neck. Motion artifacts are constant. The data is unreliable on the majority of breeds.
Some companies will continue shipping PPG because it's cheap and it checks a marketing box. Others will eventually arrive at the same conclusion we did: if you want continuous, reliable health data from a furred animal, optical sensing isn't the path.
We got there first. Not because we're smarter, but because we tested honestly, documented what failed, and made the harder choice.
Frequently Asked Questions
Why do other smart collars still use PPG optical sensors? PPG is cheap, well-documented, and has a mature supply chain from the human wearable market. It's the path of least resistance for companies prioritizing speed to market over sensor reliability. The fundamental physics limitation - that fur blocks light - remains unsolved, but PPG still checks the "heart rate monitoring" box on a product feature list.
What was wrong with PawPulse's original PPG prototypes? In our testing across 15 dog breeds and 6 cat breeds, PPG returned usable heart rate data roughly 70% of the time on short-coated breeds in still positions, dropping below 30% on medium coats and below 15% on long or double-coated breeds. Motion, ambient light, moisture, and inconsistent contact pressure all degraded signal quality further. The sensor was fundamentally incompatible with how real pets live.
How does radar solve the problems PPG has? Radar transmits electromagnetic waves that pass through fur, fabric, and surface contaminants without signal loss. It measures chest wall micro-displacement rather than light absorption, eliminating the skin-contact dependency. It operates independently of ambient light conditions and is less susceptible to gross motion artifacts because signal processing can isolate the cardiac signature from body movement.
Is radar more expensive than PPG? Yes. The radar module has a higher component cost, requires more complex signal processing, and comes from a smaller vendor ecosystem. These trade-offs increase the collar's BOM. We accepted them because a sensor that works reliably on all dogs and cats is more valuable than a cheaper sensor that fails on the majority of breeds.
Does PawPulse's radar sensor work on cats? Yes. Cats were among the most challenging subjects for PPG testing due to their dense fur and low tolerance for tight collar fit. Radar works on all cat breeds without requiring skin contact or firm pressure, making it practical for cats who reject body wraps and tight-fitting wearables.
What vitals does the radar sensor measure? The PawPulse Lucero radar sensor continuously measures heart rate, respiratory rate, and heart rate variability (HRV) from the same radar return signal. PPG, even when it worked, could only estimate heart rate - it couldn't reliably extract respiratory rate or HRV from the optical waveform on a furred animal.
Learn more about PawPulse Lucero and the radar sensing stack that makes through-fur health monitoring possible without an optical sensor.
-- The PawPulse Team
