Skip to main content

I’m (Not) Falling For This

1983

The Ambularm – an ambulation alarm

At St. Vincent Hospital, word spread quickly about the quirky new hybrid doctor/engineer down in the basement, and one of the senior physicians soon wondered if I could help him solve a big problem: patient falls. Despite instructions not to do so, patients would get out of bed and fall down on the way to the bathroom. Even the best hospitals experienced hundreds of these events per year, causing everything from bruises to hip fractures to head injuries and deaths. Bedside rails didn’t reliably prevent this, and restraining the patient in bed wasn’t acceptable either.

Weight sensors in the bed had been tried, but they didn’t activate until the patient had left the bed, and they didn’t protect a patient who was sitting in a wheelchair either. So we hit on the idea of a battery-powered tilt sensor that would be worn in the thigh, and would sound an alarm as soon as the femur (thigh bone) angled downward more than 30 degrees.

I built a crude prototype with Radio Shack parts in a plastic box with an elastic strap to hold it on the thigh. It looked promising so we had a PCB and molded case professionally designed, and eventually received a patent. The Ambularm made a distinctive bell-like sound that brought a nurse running before the patient got to his/her feet, and fall rates were cut in half. I wasn’t involved in the later stages, but apparently the Ambularm stayed  on the market until 2015.

Desperation is the Mother of Invention

1983-1985

While I was busy building up my internal medicine practice and side business in clinical software, my wife Susie’s health was deteriorating as the complications from 20 years of Type I (insulin-dependent) diabetes accumulated. Home blood glucose monitoring was becoming available, and it was hoped that more precise control of insulin dosage could forestall complications, but data management remained primitive and paper logbook-based.

In hopes of helping Susie record and visualize her blood glucose data, I added a remote terminal to my Apple II+, consisting of a TV set mounted into the wall of the kitchen and a light pen built from plans in Byte Magazine. The built-in TV made the kitchen look high-tech, and the light pen let her enter her blood glucose without using a keyboard. The software could print out a log for visits to her physician, who found the graphs printed on curly thermal paper occasionally helpful.

Despite attempts at careful glucose control, the complications accelerated, eventually leading to end-stage kidney failure, treated with at-home peritoneal dialysis. There was no invention I could come up with to overcome this setback. All I could do was help manage the thrice-daily sterile drain/refill procedures and make sure the required medical supplies were always on hand.

When severe hyperparathyroidism then developed as a complication of the renal failure, Susie underwent surgery on her neck to remove the overactive glands, but the outcome was disastrous. She was left with vocal cord paralysis requiring a permanent tracheostomy, taking away her ability to speak while recovering. Finally, this was something I could help with. I put a 555 oscillator and small speaker into a brass tube, directing the sound output through a smaller soft rubber tube. With the tube in the corner of her mouth, she could create speech with a fairly intelligible albeit robotic-sounding voice. I found the gadget still in my “junk box” 35 years later.

Complications continued to set in, and she passed away in 1985.

Quantified Self

2011-2018

After experiencing a cardiac event in 2007, I became intensely interested in the science of health behaviors and habits, and began experimenting with technology to support self-tracking and self-improvement. In 2011, I discovered Portland’s local Quantified Self meetups, eventually becoming one of the QS organizers, and giving presentations at their national and international meetings. In a QS presentation, an individual must describe their own self-tracking experiments, reporting three things:

  • What did you do?
  • How did you do it?
  • What did you learn?

This first presentation was delivered at the worldwide QS Conference in Amsterdam.

This second presentation describes my HealthESeat project, an effort to make “seat time” less harmful by encouraging me to exercise my legs while performing computer work.

HealthESeat was a “full stack” project. It included furniture modifications, proximity and rotation sensors, an LED biofeedback light, an Arduino microcontroller, and finally PC software to accumulate data and present visualizations of trends over time. Later, I added an EKG monitor for heart rate variability measurements.

I’ve found Heart Rate Variability (HRV) to be one of the most interesting physiological measurements. Research has associated it with physical health as well as psychological resilience.  I’ve experimented with ways to measure it — using either EKG or PPG biometric sensors — and ways to use it as real-time biofeedback, as well as long term tracking of trends, as described in this QS presentation from 2015.

I’ve seen benefits from my self-tracking efforts, but an even greater benefit has come from meeting such interesting people in the QS community. These associations planted the seeds for my post-retirement business, Wearable Health Labs LLC.

Show Me Some Spine!

2013-2018

My first wearable project: SpineTracker

At the Quantified Self 2012 meeting at Stanford University, I met Esther Gokhale, an expert posture teacher who had been helping people with back pain.

As I studied her research, I saw how important the posture of the spine was in her methods, and I wondered if technology could be designed to precisely measure and display a person’s posture in real time. I had training and experience in engineering and medicine, but my electronic hardware  skills were decades out-of-date, and I’d never designed a wearable device. To her great credit, Esther decided to give me a chance, and Wearable Health Labs LLC was born.

During this saga, I learned to design PC boards, have them fabricated, place tiny surface mount components using a forceps and microscope, and reflow solder the boards using a $10 hot plate and homebrew temperature controller. Then I tried some primitive Computer Aided Design (CAD) software to design prototype enclosures to be 3D printed. And finally: coding firmware for Bluetooth radio modules and software for laptops to receive that data wirelessly (I had not yet learned to write smartphone apps). After 5 years, many prototype iterations (shown in the accompanying slide deck), and considerable help from other consultants, we had a finished product. The video below shows the Spine Tracker in use.

Badges of Honor

2015-2018

A Wireless, Wearable Assist for Assisted Living

In 2015 I connected with Bill Reed and Lydia Lundberg, innovative pioneers in the assisted living industry. At their company, Elite Care, Bill had been searching for technology solutions that could satisfy the conflicting demands of safety vs. autonomy for their residents, but found no suitable commercial products. The staff and residents were already wearing identification badges, so Bill hoped we might develop these into Bluetooth wireless “smart” badges that could enhance residents’ safety and comfort, as well as staff efficiency.

Besides being nametags, the badges functioned as emergency call buttons. We experimented with auditory feedback from the call button when pressed — a short audio recording of a loved one reassuring them that help was on the way — but this didn’t prove helpful. More successful was the inclusion of accelerometer, gyro, and temperature sensors to measure activity and environment. And finally, the badges could serve as locators. With Bluetooth receiving stations (“hubs”) installed throughout the facility, a central server could estimate the location of each badge based on signal strength at each hub. The badges could also act as receivers to detect which other badges were nearby, potentially providing data on social interactions and staff presence with a resident.

Early prototype PCB

Early prototype case

Final prototype PCB

Final prototype case

Filling a KNeed

2016-2021

Wireless, Wearable Tracker for Post-Op Knee Rehabilitation

Over a million joint replacement surgeries are performed annually in the U.S., a number that’s expected to explode as my fellow boomers injure or wear out their knees and hips. While the surgery and joint implants have been continually refined, the rehabilitation phase at home hasn’t benefited from technology — yet. Orthini, a Portland startup, was formed to address this need. I came on board as a consultant to create a proof-of-concept prototype. Think of it a specialized version of a fitness tracker that measures knee range-of-motion and rehabilitation activity during the critical first few post-operative weeks at home.

Our design goals included light weight, ease of applying/removing, and no restriction of joint movement or visibility of the healing wound site. We also hoped to make it easier to apply a cold pack, and if possible, monitor the use of that as well.

Our brilliant apparel consultant, LaJean Lawson, came up with a lightweight harness that strapped to the thigh and calf, leaving the knee exposed. The electronics are hidden within the “smart buckles” that fasten the harness. This was my first experience designing an enclosure that progressed all the way to injection molding. The electronic design using a Bluetooth module and accelerometer sensors was more straightforward, but sensors to monitor the wearing of the device itself and application of the cold pack required some novel ideas. The US Patent Office agreed, finally issuing a patent in 2021 (3 years is par for that course).

Smart Buckle

KneeCoach Assembly

KneeCoach on Manikin Knee

KneeCoach Patent

Lending a Hand

2017-2018

3D Printing a Low-Cost Prosthetic Hand

I began using 3D printing for my wearable device prototypes, sending designs to a service such as Shapeways or contracting with a local firm run by biomedical engineer Rachel Dreilinger. But I wanted to iterate on my designs more quickly and at lower cost, so in 2017 I built my own 3D printer from the Prusa i3 MK3 kit.

I discovered there’s a learning curve to 3D printing, but fortunately there’s also a massive online community of “makers” adopting the technology — over 1000 right here in Portland! –so there has been plenty of help available. One community, e-Nable, has tackled the challenge of 3D printing free prosthetic hands for clients in lower medical resource environments. It’s well organized, with an informal certification program for fabricators, and the accompanying video was my application for approval.

Designing and Printing a Robotic Gripper

As 3D printing became easier, public libraries stepped in to provide access to this technology. The Hillsboro Public Library built a well-equipped “Collaboratory” makerspace where I served as a volunteer, teaching a class and helping patrons use the devices available. A fellow volunteer had taken on an interesting project — designing and fabricating a sophisticated robotic arm — and given the daunting size of the project, he welcomed my offer to design a gripper to go at the end of the robotic arm. You can learn about my efforts in the associated video here.

Hearts on Fire

Cardiac Stress Monitoring and Notification for First Responders

2017-2020

After retiring from full-time work at CCHIT, I accepted a volunteer role closer to home, as an Advisory Board member for the Berglund Center at Pacific University. Following several years serving as a judge on their annual Inspired Ideas competition, I was asked in 2017 if I could take on a bigger role: mentoring their winning team of students to guide them from their raw concept to a working prototype.

The team of psychology graduate students had been studying and treating stress in firefighters. They, and I, were surprised to learn that heart attacks — not fire, smoke, structure collapse or falls — were the leading cause of on-the-job deaths in firefighters. The students wanted to equip firefighters with heart monitors that would measure and relay their stress levels to the incident chief, but lacked any engineering training or resources. I liked this challenge immediately and accepted a role as their engineering consultant.

Wearable physiologic monitors were an already well-established technology, appearing in smartwatches and fitness bands. But there was a showstopper: their low power 2.4 GHz Bluetooth transceivers could not achieve the communication range and reliability needed at an incident site, and we could not rely on a consumer smartphone as a relay in this demanding application. Then the proverbial light bulb of invention clicked on: every firefighter already had a robust wireless communication device: a UHF portable radio, controlled by a shoulder-worn speaker/microphone. We conceived of a “smarter” speaker/microphone unit that included a microprocessor and Bluetooth transceiver to receive the physiologic data. Heart-rate based alerts could then trigger a synthesized voice message. The alert could be played locally over the speaker to notify the firefighter him/herself, and then — if more urgent — it could activate the push-to-talk circuit and be transmitted over the portable radio’s network to the incident commander.

I started with a breadboard consisting of an Arduino-compatible microcontroller, an MP3 player, a Bluetooth module, and relays to trigger the PTT and switch the audio. The students’ jaws collectively dropped when they saw their idea actually working, despite the ugly breadboard implementation. The next step was to shrink the size by designing a custom PCB and a 3D-printed case, creating a first prototype that could at least fit in your hand. For the next iteration — now called HeartMic — I plunged into C programming to get all the necessary code running on the microprocessor included in the Bluetooth chip. With the separate microcontroller eliminated, the device was slimmed down to fit in the palm of one’s hand. The finishing touch was laptop software, communicating via Bluetooth, to configure HeartMic with specific alert levels for each firefighter. We demonstrated the system at a fire station in Washington state, and applied for a patent which was granted in 2021. The students truly had a “full stack” product development experience!

Shrinking Circuits, Meet Shaking Hands

DIY assistive technology for assembling miniaturized electronic prototypes

2020

When I first started tinkering with electronics in the mid-20th century, most circuits used vacuum tubes. Over the 60 years since then, I’ve been privileged to witness the stunning progress in miniaturization, first with transistors, then basic integrated circuits, all the way to highly complex chips the size of your thumbnail with billions of transistors inside. To keep pace, small passive components — resistors and capacitors — shrank too. To build a prototype of a wearable device with a small size and weight, you have to be able to assemble these minuscule chips and components.

To assemble a circuit, I apply solder paste to pads on a printed circuit board using a laser-cut stencil. A boom microscope provides adequate magnification to see the parts, but the final step requires using a hand-held forceps to pick up and place the tiny components in exactly the right positions on the PC board. By the time I reached 70 years of age, the parts had shrunk to 0.040 x 0.020 inches size, with placement tolerances of 5 thousandths of an inch. And I found I could no longer grip, maneuver, and release the part from the forceps with the rock-steadiness required. Unwilling to give up, I hatched a DIY solution that restored my ability to place components at this new level of precision. More details are available after the photographs.

The first problem was one of crosstalk between two manual processes: moving the part into position over the PC board, and releasing pressure on the forceps at the right moment. The solution was to use a vacuum pick-up device instead of a forceps, and to reassign the function of releasing the vacuum to a foot pedal instead of my hand. From the photos, you’ll see how I started with an inexpensive vacuum pump, then added solenoid valves and associated circuitry to allow foot-pedal control of the vacuum.

The second problem was that of tremor. With aging some neurons conduct impulses more slowly or even stop functioning, and a servomechanism with excessive delay in its control loop will overshoot and oscillate — and that causes tremor. The solution was to add a stabilizing fixture, with a fulcrum close to the PC board, and damping to reduce the oscillations. The stabilizing fixture was 3D printed, and its fulcrum area and the pickup tool were covered with rubber tubing selected for its damping characteristics.

Sensors for Seniors

2020-2022

Pandemic Project #1: Wireless sensors and gateway for CareBank

The Covid-19 pandemic was devastating to public health, but it also stimulated the development of new technologies hoping to re-establish our social connections. I met Claude Goodman in March 2020, learned of his CareBank project, and gladly took on the challenge of developing improved wireless sensors for that system.

The concept of a Bluetooth Low Energy module and motion sensor was not unusual, but long-range transmission, long battery life, and ease of battery replacement by an elder were more demanding requirements. We used BLE 5.1 and careful antenna design to extend the range and a low data rate to achieve >1yr of battery life. To change the battery, the user just slips off a silicone band (blue band in the rendering) and slides a new coin cell battery into a slot.

The next challenge was reliably relaying the sensor data to the CareBank servers in the cloud. While an up-to-date smartphone with an Internet data connection could do that job, it wasn’t realistic to assume every elder had such a smartphone always operational. As an alternative, I developed a Bluetooth-5-to-cellular gateway prototype. To maximize range, antenna placement was carefully optimized, and the communication link uses positive acknowledgement and retransmission to overcome dropouts. The gateway’s 3D-printed enclosure plugs into a standard USB charging brick, and the combination plugs into a wall outlet like a nightlight.

When it was time to manufacture a prototype run of 75 sensors, we hit one more challenge: the 2021 global semiconductor shortage. I learned how to source components from dwindling stocks in multiple countries and have the PCBs fabricated and assembled overseas. We were lucky; some key chips now have a one year lead time.