KINETIC: Intel Edison, Gravity, Health and Economy

Kinetic: Intel® Edison Compute Module Powers Wearable Protection Against Lifting InjuriesTranslation: Intel blog, Geektimes

The surest way to protect yourself from back injuries when working with weights is to give such work to someone else. But seriously, the correct handling of heavy objects is a whole science. For example, many people know that it is necessary to lift and lower weights with a straight back, so that the main load falls on the legs. Deviation from only one of these rules will sooner or later make itself felt, and there are others.

Violation of safety at work turns into injuries to employees and losses of companies. In 2013, for example, back injuries cost companies in the US $70 million. There is a problem, unpleasant and expensive. A startup from New York, KINETIC, which deals with intelligent wearable devices, has found a way to use Intel Edison computing modules to create a system that can significantly reduce injuries in dangerous workplaces and save companies huge amounts of money.

KINETIC uses sensors in a device that is attached to a belt, as well as well-thought-out algorithms that help determine whether a person is working with weights safely, or so that his back is about to break. An additional module resembling a wristwatch and equipped with a vibration motor is able to instantly notify an employee that he is committing dangerous actions. For continuous professional development of employees, which means work safety, at the end of the shift, data from devices is collected, uploaded to a cloud service and analyzed. As a result, this data turns into valuable analytical information, into statistical information that can be viewed through the web interface. This analytics can be used, for example, for additional training of employees and in order to change something in the workplace and make them safer.

How could a small startup, which is only a year and a half old, be able to quickly develop an industrial-grade IoT device, write programs for it and bring it to market? Today, together with the founders of KINETIC, we will tell the exciting and inspiring story of this project.

How it all started

The CEO and one of the founders of KINETIC, Haytham Elhavary, was educated in the field of biomedical engineering. He wanted to make sure that people who work with weights suffer less from injuries in the workplace. A similar desire was aroused in him by memories of how his mother, who worked as a nurse, constantly complained of back pain. This is not surprising: she had to help patients, often completely helpless. Here's what he says about the project: "I wanted to take technologies that have been available on the market for a long time and introduce them to workplaces, where these technologies can help in reducing industrial injuries."

He saw movers as "industrial athletes", so he found something similar to what he wanted to create among sports devices, such as Nike Plus sneakers. They are equipped with sensors capable of recognizing movements, the direction of movement, acceleration, ascents. Based on the collected data, it is possible to determine exactly what a person is doing, calculate the distance traveled, and give him recommendations.

"As is the case with any wearable device, or with something that relates to the Internet of Things," says Haytham, "the project has a hardware component that requires special knowledge in design and production. Then there are sensors that collect data. In order to extract something useful from this data, you need not only knowledge of how to interpret them, but also the ability to clear them of noise and highlight the most important thing. The next stage is the development of the user part of the project. In particular, an information panel with clear analytical information."

In order to give life to the project, Haytham and another founder of the project, as well as his CTO, Aditya Bensel, an electronics engineer with experience in developing wearable devices, enlisted the support of a data processing specialist, a software engineer and two mechanical engineers, one of whom had manufacturing experience. Let's see what they got.

Getting Started with the Intel Edison Computing Module

The ability to collect data on the actions of workers and give them instant feedback when they handle heavy loads, risking their health, means that the device must perform calculations in real time.

"This is how some processes are usually automated in warehouses or factories using portable devices. First they collect data, then they send it to the cloud, perform the necessary calculations there. Then they transmit the results back. The devices, in turn, somehow react to the accepted results," says Haytham.

This idea, when applied to KINETIC goals, has a serious drawback. The fact is that not all areas of warehouses or factories can be connected to Wi-Fi networks. The KINETIC team needed a solution that could process data locally using the hardware capabilities of a portable device. "The Intel Edison computing module played a crucial role in our project," says Haytham. "It is very compact and at the same time has excellent computing capabilities. Edison is able to process data immediately after they arrive, instantly respond to dangerous situations without relying on external services to which you need to connect via Wi-Fi."

Intel Edison, a miniature, energy-efficient device, has a built-in dual-band Wi-Fi adapter. Edison supports Bluetooth 4.0, is equipped with a dual-core Atom processor and an Intel Quark microcontroller. Haytham believes that Edison has given KINETIC an excellent multi-purpose platform, and this, among other things, means flexibility of the solution. What kind of "flexibility" are we talking about? Here is how the founder of the startup answers this question: "Now we are busy protecting against injuries when working with heavy loads, since they account for the lion's share of compensation claims from employees. As we develop, we want our solutions to be used to analyze other types of work, the incorrect execution of which is traumatic."

KINETIC's wearable device relies on Intel's Edison computing module to work with its sensors, collecting data and storing it in flash memory. The device uses the built-in Edison Wi-Fi adapter to upload data to the cloud. This is done at the end of the shift. Bluetooth is used when additional wearable sensors are added to the solution. It is with the help of BT that they interact with the main module.

Anatomy of devices from KINETIC

The system, which was developed by KINETIC, includes two devices. The first, the main one, is worn on the belt. It, among other things, contains sensors and a computing module. The second, optional, resembles a wristwatch. It is worn on the wrist, it contains additional sensors and a display. When it is used, both devices work together, collecting more accurate information about the actions of workers.

KINETIC's smart wearable devices protect workers from back injuries by identifying potentially dangerous activities in real time.

The device, which is worn on a belt, includes a built-in accelerometer and measures acceleration, tilt angle, position in space using the company's own algorithm. "We observe the movements that an employee makes when dealing with weights and loading the spine," says Haytham.

A device from KINETIC, which is attached to the belt. It is capable of operating autonomously and includes sensors and an Intel Edison computing module.

Modeling of actions

Haytham's experience in biomechanics has played a key role in the development of algorithms for monitoring employee activity and determining how safely certain actions are performed.

"A lot has been written about this," says Haytham. "In the last 20 years, scientists have developed and published models that describe indicators that are indicators of risk. We found out which indicators work best."

Instead of reinventing the wheel, the KINETIC team automated the best they could find. "Ergonomists usually do this: they stop a person in the process of performing some operation, say, when lifting a heavy box, and, literally, they begin to measure, for example, the bend of the back and knees, and how far the box is from the lower back," says Haytham.

For research purposes, all this is very useful, but in the real world, in production, for example, such "manual" measurements are not applicable. According to the founder of the company, thanks to KINETIC, for the first time in history, it became possible to automate such measurements, and if the risk is too great, immediately notify the employee about it.

How does KINETIC find out, for example, that a certain box is held too far from the human body? "An additional wrist module is used to solve this problem," explains Haytham. "The device attached to the belt can track the movements and bending of the back. Only with its help you can do very, very much. However, if more detailed data is needed, for example, you need to find out how far the box is from the waist, or if you need to get data on lifting weights above shoulder level, which is considered a very risky business, then you need a module that is attached to the wrist."

Behavior change

KINETIC spent a lot of effort on creating a feedback mechanism, thanks to which employees learn that they are doing something dangerous. "Managers would really like the idea of a taser. But such feedback will definitely not please employees," Haytham laughs. "And no kidding, the device on your wrist vibrates like a phone vibrate when a message arrives. The main purpose of feedback is to change dangerous behavior scenarios to safe ones."

Adding to the difficulties is the fact that employees, even at the expense of safety, are trying to do as much as possible. Much here depends on the forms and systems of remuneration. According to Haytham, people are trained to work safely, but often, without paying attention to what they were taught, they choose the fastest ways to perform certain actions.

KINETIC wanted to remind employees of what they had been taught. In the summer, the company launched a pilot program in several warehouses of companies that work in the areas of cargo delivery and logistics. This program gave them some important information. "Alerting people about dangerous actions, in general, turned out to be useful," says Haytham. "And in order to enhance this effect, employees need to be motivated, they need to be allowed to interact with the system, make it understandable."

The KINETIC device is quite thin, which makes it convenient to wear it on your belt

Older generation workers are usually attentive to occupational safety and health, while younger and stronger people think they are invulnerable. In order to motivate them all, KINETIC uses a method well-known in sports: competition. Namely, a record is kept of actions performed with a high risk to health, which each of the employees performed during the day. Then the results of the workers are compared. The one who has performed the most high-risk operations safely receives a small reward.

KINETIC borrowed another idea from the field of consumer wearable electronics: the leaderboard. It allows employees to see what their results are compared to previous working days and compare their achievements with the indicators of other employees. KINETIC says that people were attracted to this idea, they started discussing it with each other: "I can't believe you got ahead," "We need to get around it." The work turned from a boring sequence of actions into an exciting competition.

Checking the leaderboard

In addition, this approach helped to build trusting relationships in the team. "When we gave employees access to the leaderboard," KINETIC says, "they were able to freely see their own and others' results. They realized that they shouldn't worry about something happening in the depths of our system that they don't understand, but can somehow influence them."

Data processing

Sensors collect data, but, along with a useful signal, there is a lot of noise in this data. "The data from the motion sensors is very noisy," says Aditya Bansal. "We use a lot of traditional signal processing algorithms. The computing power of Edison proved to be very useful for solving such problems."

On the left – noisy sensor indicators. According to such data, it is difficult to understand how, in fact, a person moves. After processing the signal, it becomes clearly visible when exactly the person bends down (the blue line goes up) and returns to the starting position (the blue line goes down).

A special kind of noise is the actions of the employee himself. Such a "noise" occurs when a person, for example, is having lunch, waving his hands or doing anything that does not relate to working with weights. "This brings us back to what Haytham has already touched on, talking about algorithms for processing sensor readings and finding out exactly what an employee is doing," says Aditya. "We need to drop everything that is not related to working with weights."

In order to solve this problem, the KINETIC team recorded the activities of employees on video. "We compared the information from the sensors with what was on the video. This was done in order to find out exactly what people are doing and how sensors react to it," Aditya explained.

Connecting to the cloud

At the end of the shift, workers install their devices in the charging station. At the same time, all these devices begin to transfer data to the cloud. The project uses the Amazon Web Services (AWS) API, which allows you to securely transfer information directly to the cloud service.

The Amazon Elastic MapReduce (EMR) analytical system processes a huge amount of information collected from hundreds of devices in real time and stores them in a database hosted in AWS. After that, the data enters the KINETIC dashboard, giving managers meaningful indicators suitable for further work with them.

Generalization of experience

Given the huge amount of data collected, it is not surprising that KINETIC has discovered some interesting patterns in employee behavior. "We saw clear bursts of dangerous activity at certain points of the working day," Haytham said. "At one plant, this happened right before the lunch break and shortly before the completion of work. We think this means that people are trying to do more before the break or before the end of the shift."

"Imagine that you are the head of the security service," he continued. "You are collecting data on high-risk actions performed by employees of the company. Then you decide to implement a new training program or spend money on more convenient and safe equipment of workplaces. After the introduction of innovations, you should see a change in indicators. If there is no difference, then what you have done has not changed anything. This allows us to assess the real impact of measures to improve security on the situation. Until now, the only way to understand whether the innovation is working or not has been to simply wait and evaluate the change in the number of injuries."

Contribution to the community: seminars

Haytham Elhavary and Aditya Bensel are actively involved in the life of the developer community, exploring new ideas, supporting talented newcomers and helping others bring their developments to market. They do this by organizing monthly seminars in New York.

"We talk about the experience gained, about how to bring the product to market, about building a brand. We pay attention to production," says Haytham. These events attract more than 200 people every month, and the list of members includes about 3,000 developers. On this occasion , Haytham remarks: "I think this indicates the growing popularity of hardware startups."

Aditya agrees with this: "I conduct seminars on embedded systems here in New York. They are less focused on business. These seminars are mainly devoted to software and the technical community. Many companies, including Intel, are represented at the events. This speaks to the size of the community formed around embedded systems in New York."

Results

Intel has made a serious contribution to the IoT community. This is, for example, the support that the company provides to those who are interested in Intel Edison. KINETIC is one example. Intel's participation played an important role in their project, helped to quickly develop an intelligent wearable device connected to the cloud, and put it on the market. The assistance, in particular, consisted of answers to technical questions and support in the field of the program part of the project. The Intel Edison computing module turned out to be an ideal choice in the field of local data collection and processing and securely sending them to the AWS cloud for analysis.

KINETIC actively participates in Intel Development Zone forums "Whenever we encounter a problem and find a solution to it, we immediately publish an answer," the company says.

KINETIC fulfills the noble mission of improving occupational safety and reducing costs. The company, without exaggeration, protects the "backbone of the economy". KINETIC has paved the way for IoT solutions based on Intel Edison computing modules, sensors and AWS cloud services to places where helmets, metal-toed safety boots and safety glasses have been "security technologies" until now.

Portal "Eternal youth" http://vechnayamolodost.ru  14.07.2016