3D TIME OF FLIGHT TECHNOLOGY ENABLES TEMPO AT-HOME PERSONALIZED FITNESS TRAINING
During the COVID-19 pandemic, people isolated themselves to stay safe and healthy, eliminating many forms of social contact, including in person workouts at gyms. For many of those people, that meant curtailing their workouts altogether, which led to weight gain and decreased overall health and wellness.
Technology, innovation, consumer demand, and a void in the market converged to generate significant demand and market opportunity for at-home gyms, designed to help people stay healthy and shed those extra pounds. However, given that humans are social creatures, many still crave one-on-one interaction, and yearn for personalized instruction and motivation.
Tempo recognized the need for personalized training and guidance and set out to develop Tempo Studio, the first at-home gym with a built-in personal trainer. The complete workout system required a 3D time of flight (ToF) depth recognition technology capable of measuring a subject’s position and tracking their rapid movements. To meet the challenge, Tempo teamed up with fitness scientists, physiology PhDs, and certified personal trainers to deliver AI-powered personal guidance, real-time feedback, and custom training plans and programs.
Tempo collaborated with Analog Devices, Inc. (ADI) to deliver greater subject position tracking via ADI’s high resolution 1 megapixel depth sensing ToF technology. The high accuracy, low latency solution enables enhanced feedback for a more effective workout for fitness enthusiasts across the globe.
AT A GLANCE
COMPANY
Founded on the philosophy that personal training changes lives, Tempo is redefining the at-home fitness industry through a growing suite of AI-powered advanced technology products and bringing effective, results driven models to more people at home.
APPLICATIONS
3D Tempo Vision™ leverages ADI’s industry-leading ToF technology system and sensors, enabling AI-powered personal workout guidance, real-time form feedback, and custom training plans delivered via the internet.
CHALLENGE
Bring to market the industry’s first 3D, high resolution, high frame rate, at-home personalized health and fitness system. Measure and track the exerciser’s position and rapid movements with increased depth measurement and accuracy at an affordable price.
GOAL
Improve and maintain the health and fitness of at-home boxers, meditation practitioners, strength trainers, and yoga enthusiasts with a safe and convenient personal trainer system that offers explicit guidelines and corrective measures.
THE OPPORTUNITY
ENTER TEMPO
Tempo’s entrance into the fitness industry began with personalized training systems in gyms, followed by at-home products. After building a reputation with health professionals and athletes such as ski legend Lyndsey Vonn, Tempo formed a collaboration with Microsoft to leverage the technology giant’s Kinect camera platform.
Microsoft’s Kinect is a finished RGB camera capable of motion sensing and measuring depth using ToF infrared projectors and detectors. However, Kinect was not designed for high volume production or embedding the depth camera into other systems. A new optimized design was needed—one with efficient systems integration capability.
*The Business Research Company’s Online/Virtual Fitness Global Market Report 2022
“Microsoft licensed ADI’s time of flight core technology to help proliferate 3D image sensing beyond their Kinect platform,” said Tony Zarola, Senior Director for Enhanced Imaging and Interpretation at ADI. Microsoft wanted to create more 3D cameras that could leverage its Azure Cloud Services, thereby driving more business.
Tempo migrated away from Kinect to develop its next-generation fitness solution. Tempo developed their own 3D camera based on ADI’s ToF underlying technology—featuring high resolution 1 megapixel depth sensing.
THE ENGAGEMENT
Since 2019, ADI has supported Tempo at its factory with domain and technical expertise. Tempo initially planned to build its own ToF module, however, after collaborating with ADI’s ToF team, Tempo decided to use ADI’s 2023 CES Innovation Award-winning ADTF3175 ToF module to miniaturize the whole system and reduce complexity, power, and size while maintaining the performance of Kinect.
“We are grateful for ADI’s technical expertise and support during the product development process,” said Kailash Hiremath, VP of Hardware at Tempo. “ADI’s technology will enable us to bring to market the next generation of Tempo 3D Vision and enable us to deliver a market-leading fitness experience for our customers.”
THE CHALLENGE
The Tempo personalized fitness training solution needed the ability to track and distinguish these essential elements.
- Distance of the equipment and individual from the ToF camera
- Precise position and rapid movement of the person exercising (arms, legs, torso, head)
- Precise position and rapid movement of the fitness equipment
Determining distance was the first order of business.
TIME OF FLIGHT 101: MEASURING DISTANCE
A ToF camera measures distance by bouncing a beam of light off an object and back to a sensor. The ToF camera then measures the time delay between when the light is emitted and when the sensor receives the reflected light. The process is similar to ultrasound, which uses sound to measure distance, and radar, which uses radio waves. ToF cameras can generate high resolution depth maps with precision depth accuracies and spatial resolutions similar to RGB cameras. While radar has a longer range, ToF is more accurate, with higher resolution.
FAST, ACCURATE, AND DETAILED TIME OF FLIGHT DETECTION
Tempo demanded high accuracy distance measurement and the ability to rapidly generate enough detailed information to capture a bar a few centimeters wide, in motion most of the time, and know precisely where it is in real time. ADI’s 3D ToF technology met the need with high resolution, low latency, high frame rate, wide field of view, and extended depth.
High Resolution
ADI’s CMOS imaging chip generates an unequaled million pixels resolution of a training scene. The result is sharp edges, clarity between the different elements, and increased depth information to determine where, for example, the barbell is and where the subject’s hands are relative to other parts of the body. This fine level of detail enables the AI trainer to determine if the subject is performing the exercises correctly.
Depth Processing
There are two ways to create depth in a scene. The first is by using software. The second is through sampling. Tempo initially approached the task using the former but found that power, low latency, and a high frame rate were compromised when running the depth compute in software. They subsequently took advantage of ADI’s new technology—depth computation via sampling on a chip. The ADSD3500 component reduces system power significantly and enables customers to achieve low latency and high frame rate operation.
3D Camera
ADI’s 3D, continuous wave (CW) CMOS ToF cameras provide the highest performance solutions on the market with the most reliable and precise depth information, even in large spaces with highly reflective surfaces and moving objects.
High Frame Rate
ToF technology is capable of achieving 40 frames per second. The high frame rate aligns with Tempo’s capacity and helps to eliminate blur and pinpoint the position of various elements as users move in the scene.
WHY HIGH PERFORMANCE 3D TIME OF FLIGHT IS CRITICAL
With a less powerful ToF system, one with lower resolution and frame rate, you might not be able to see an exercise bar or its positioning because of its speed of motion. Distance from the camera may also be a factor, as some systems are limited in range.
A SMARTER HOME GYM
It’s a hardware, software, and content streaming device that packages an immersive 42" HD touchscreen display, with premium weights and equipment, all in an award winning free standing design.
“We’re adding another human sense—VISION!”Tempo |
With Tempo, your form is tracked and analyzed by built-in virtual trainers who can see your movements and help you adjust your form in real time. Tempo’s AI brain learns your body and goals, offers personalized weight recommendations, and delivers guidance and pointers so your progress is faster. It’s the only smart gym with AI-powered training built into every class.
“Tempo is the first and only at-home fitness system that can track your motion, count your reps, and improve your form. Using 3D sensors and AI, Tempo guides and offers you personalized feedback while providing a richer, more effective, and safer workout,” said Kailash Hiremath. “The road to health is personal, and the core aspect of our products is around individualized feedback.”
“ADI helped us to redefine the at-home personalized fitness industry with its performance leading time of flight technology, enabling our 3D Tempo Vision that powers our studio product.”
ADVANCING INDUSTRIES
Like the 19th-century steam engine—a catalyst for the industrial revolution—ToF and AI are advancing and redefining many of the world’s industries.
From navigation for autonomous manufacturing to industrial machine vision for quality inspection, the manufacturing industry is adopting ToF sensing technologies. It’s also becoming a vital part of consumer electronics from augmented reality (AR) and virtual reality (VR) headsets to smartphones with advanced photographic and security features. Already on the drawing boards are cabin ToF automotive systems that will monitor the position and state of all the people in the vehicle and allow drivers to adjust climate control through a simple hand gesture. Lifesaving ToF applications already permeate healthcare, saving lives and ensuring health with the remote monitoring of infant respiration and social distancing in hospitals and clinics.
Leading the way to ToF and AI innovation, ADI offers visionaries like Tempo the highest performance 3D ToF—cameras with 1 megapixel resolution CMOS imagers, software and hardware-based depth compute and processing, laser drivers, and power management. ADI also provides 3D ToF platform tools and kits for software and algorithm development and prototyping (EVAL-ADTF3175D-NXZ), allowing for more rapid evaluation and implementation.
