As an engineering partner for leading industry startups and software product organizations, Virtusa can provide a unique perspective into what lies under the hood of many of today’s most disruptive technology platforms. The Virtusa Garage Series focuses on taking a deeper dive into explaining both the technology and broader business models behind these innovative companies.
It seems like everything is getting so incredibly smart these days: Smart Cities, Smart Energy, Smart Factories, Smart Homes, Smart Cars, Smart Medicine. Even our kids constantly remind us of how much smarter they are than the generations of the past. There is just no getting around it. Society is moving to be a much smarter place and we are all moving a bit closer to living Jetson-like lifestyles. For me, I knew that we had come at a true crossroad when our family sewing machine got connected to the Internet.
Clearly there is a lot of technology driving this change but the one technology which may be underlying the most dramatic ‘futurism‘ shift is the Internet of Things (IoT). It may be hard to refute that the Internet of Things is by far the coolest of all names given to any of our technology movements. Several years back, IoT was hyped as much as Blockchain is today. Everyone was talking about IoT as the next technology phenomenon which would truly change the way we ran our lives and businesses. Endless numbers of sensors would collect data on everything we do and help to automate the world. More than half the world’s data would soon come from edge-based devices. Then things quieted down on the IoT hype. Or did they really?
Two market segments have formed: the consumer market and the industrial market. The latter and much larger industrial market is forming into its own subsegment of IoT called the Industrial Internet of Things (IIoT).
In the consumer market, we see IoT devices around our everyday life. Smart watches, smart home devices, smart cars, and smart medicine to highlight a few. Taking pills with embedded chips to help diagnose medical conditions brings back memories of the early sci-fi flick Fantastic Voyage. Asking Alexa or Siri to adjust our home’s temperature, close/open a garage door, turn off/on lights (without a clapper), vacuum the floor and even order groceries when the fridge is running low seems astonishing. And let’s not forget that we can now even download the latest quilting patterns onto our sewing machines. It becomes hard to imagine how we survived without this amazing stuff in the past. Especially for the dwindling few who remember rotary phone dialing and rabbit ears on our RCA televisions.
On the industrial side, some of the most notable progressions include smart cities, smart energy grids, smart manufacturing, and connected health. Not to take the words out of our President’s mouth, but the global IIoT growth expectations are simply ‘Huge’. IIoT global spend estimates are projected to top $1 Trillion dollars by 2025. Yes, that is a Trillion – with a Capital “T”. Major companies including GE (Predix), Siemens (Mindsphere), Hitachi (Lumada), Schneider Electronic (Wonderware), Cisco (Kinetic) and IBM (Bluemix) have all invested heavily in creating their own IIoT platforms. All major cloud providers (AWS, Google GCP, Microsoft Azure) and big tech giants (IBM, SAP, Oracle) have embedded IoT frameworks into their core operating environments.
Just too much capital is being invested into our global technology backbone for IoT not to drive significant change into the way we fundamentally work and live our lives. IIoT is anticipated to improve productivity, enhance worker safety, reduce operating costs and spawn off new business verticals. To gain the highest level of efficiencies, our workforce will be blended, and it is expected that humans and machines will work together to achieve results that humans nor machines can produce alone. The good news is that we all saw the Terminator movies and it all works out in the end; after of course a few speedbumps in the early part of the human-machine teething process.
With this as the backdrop, the Modjoul platform was built to help organizations address many of these challenges with a heterogeneous IoT-based platform that can integrate worker data coming from wearables (e.g., SmartBelt) with machine-related data (e.g., forklift, truck, or factory) and other structured or unstructured corporate data. Modjoul provides the ability to correlate data from multiple sources into a common AI under-laced data model. Combining Information Technology (IT) and Operational Technology (OT) into one ‘Open’ platform with ease of use APIs to mesh everything together. It’s innovative companies like Modjoul who will help pave the way for IoT to eventually have an unparalleled impact onto our society.
Let’s take a deeper dive into both the Modjoul company and their underlying platform.
Co-founded in 2016 by Eric Martinez and Jen Thorson, Modjoul provides a comprehensive Internet of Things (IoT) data aggregation and analytics platform to help organizations proactively manage worker safety, improve organizational efficiency, reduce operational risk, and much more.
As the former Chief Claims Officer for AIG, Eric logically targeted his original vision for the platform’s initial distribution at the insurance carrier market. Insurance is an industry that has been an early adopter of IoT/telematics-based business models to help better manage insurance risk as well as introduce new insurance product offerings. Examples include usage-based (UBI) and pay-as-you-drive (PAYD) programs. UBI and PAYD plans allow individual policyholders and commercial fleet operators to track driving habits and share this information with their insurance provider for better rates.
Modjoul’s original business model followed a similar telematics use case but focused on collecting data from workers rather than machines. By providing industrialized wearables to workers, a company could proactively manage worker safety while also collecting a worker’s activity data. This data could also be useful in determining the cause and validity of employee disability claims. This data, in time, will inform insurance carriers on the actual risks across industries and various job types.
Injuries in the workplace are occurring without clear or practical resolution. In 2016, there was an estimated 2.9 million workers injured at work and 5,200 worker fatalities. The work done by the employees is generally not being tracked, analyzed, or used in a meaningful manner. Furthermore, collecting and analyzing work metrics is a time-consuming, error-prone task. The Modjoul SmartBelt enhances worker safety and productivity with a mobile app and an intelligent workplace wearable using a cognitive IoT platform to provide a well-connected work environment.
As with many business models, they quickly evolve as they are taken to market. In this case, while the insurance market still has tremendous downstream adoption potential, the majority of the client base for Modjoul has been large corporate clients across multiple industries with transportation and logistics, manufacturing, and agriculture being the early adopters. Leading Fortune 500 companies have been quick to embrace and adopt the platform. Many of whom had already significant investments being made in IoT. What Modjoul provides is to some extent the missing link. Adding detailed worker data into a platform which can aggregate in other machine and corporate data to form a comprehensive data set to help drive a higher degree of overall safety, productivity and efficiency through intelligent automation.
SmartBelt Solution Overview
The Modjoul SmartBelt is the company’s first wearable device delivered to market. Think of it as an industrial Fitbit that can capture significantly more data than many of today’s consumer-oriented devices. The sensors in the SmartBelt collect various data activities from the user to determine their safety while on the job. By collecting data such as location, motion, and environment, companies can have a comprehensive view of how and in what conditions their employees operate. Equipped with sensors and GPS, the SmartBelt provides new data to help companies understand behaviors and activities that often lead to injuries. Data is collected based on counts, duration, and rate of the employee’s activities. The SmartBelt can pick up multiple movements and work motions, including walking, sitting, falling, driving, tripping, and more. As the data aggregates, the SmartBelt simultaneously converts the data into working “verbs” and transmits the information to the cloud.
Once the data is collected, orchestrated, and classified, the information is reported in easy-to-understand metrics on three customizable dashboard views: Organization Administrator, Supervisor, and Employee. The dashboard reports are easy to access via the Modjoul app or online and allow the viewer to see an overview and ranking of an employee’s safety. It’s also possible to drill down in the dashboard to look at specific events and to track an employee’s location through the breadcrumb report.
The use cases for such a technology seem endless. Are my crop pickers taking enough water breaks? Are my forklift drivers keeping themselves out of potentially harmful mischief? Are my factory workers staying clear from restricted hazardous zones? Is my delivery workforce using the correct lifting techniques to help prevent undue back and muscle strain? Are the packages actually getting delivered to the designated address? Is there a man down situation with my first responder workforce?
Solution Deep Dive
Now comes the interesting part. What is the Modjoul platform and why does this company have the potential to succeed where others have struggled to really hit scale in this market? Well, we know they have a SmartBelt, one that was engineered in partnership with some of the best engineering minds at Clemson University. But the belt is just a unique wearable within a much more comprehensive IoT platform. It is an entirely API-driven platform and is tremendously scalable and extensible. Through APIs, data can be streamed from any wearable device or machine, including belts, vests, and hard hats. The connected workplace can be extended to phones, chip-enabled ID badges, forklifts, trucks, and factory machinery.
The back-end architecture of Modjoul is where the heavy lifting takes place. Data is securely received in millions of bits and is intelligently reassembled. The back-end application platform provides an advanced framework for a company to monitor data intelligently and receive alerts based upon established thresholds built using machine learning algorithms. In addition to the application platform provided, Modjoul has an open API architecture so organizations can also move the data to their own environments for advanced analytics and third-party device integrations.
The SmartBelt contains numerous components and sensors including accelerometers, gyroscopes, pressure sensors, a battery, and GPS to turn the human body into a vast location-specific data generator. TLS-encrypted sensor data from the SmartBelt is continuously streamed from the SmartBelt to the IoT bridge cloud in a secure fashion. The bridge streams the messages through an IoT gateway into a Kinesis stream that in turn is consumed by a Spark streaming application. This formats the data into key-value pairs and enriches the payloads with data from CouchDB.
The enriched data is then written to HDFS in parquet format to enhance performance. Near-real-time jobs classify and transform data for loading into CouchDB. CouchDB auto-indexes the documents and aggregates data for consumption by trained machine learning models that enable intelligent monitoring of the workforce.
Data is parsed through pattern recognition and data modeling and is converted into simple verbs—think walking, driving, bending, climbing, and sitting/standing (idle). Modjoul currently employs two random forest models within the batch processing. The first model is a high-level activity classifier and provides an activity label for each millisecond of data (e.g., static, dynamic, upper-body, or lower-body movement). This classification is included as an input (feature) in the second model, which provides lower-level classification for each second of movement (e.g., walking, sitting, or twisting). Beyond this, the platform employs many heuristics to arrive at other classifications like driving, acceleration, hard braking, and squatting that lead to intelligent Gait analysis—more specifically, the study of human motion—using the SmartBelt to augment the instrumentation for measuring body movements, body mechanics, and physical activity at a particular geolocation.
All jobs resources are managed by Apache Yarn and scheduled through an Oozie workflow. Bisynchronous messaging sends alert messages to the edge device. If a worker is sensed as being in a fatigued state while driving or perhaps moving too close to a factory danger zone, the belt can provide a warning vibration. Processing happens 24×7 and near-real-time streaming data is reflected on the dashboard. AWS’s elastic scaling features are leveraged to distribute workloads into lower-cost time windows.
The platform has been architected to support multiple business use cases and is already being extended to support additional wearable devices. Built as an API driven architecture, Modjoul is also offering its encapsulated data models in a subscription based model on Amazon’s marketplace. Examples include; a Walking Model, an Asset Utilization Model, a Lower Lumbar Model, a Geo Fence Model, Motion and Stationary Models with many more to come.
For organizations looking to truly automate their overall operations through intelligent use of IoT data, Modjoul is certainly worthy of a detailed evaluation. It is built as an industrial-strength platform that can easily scale on the back end of a flexible AWS cloud platform or, for larger organizations, a hybrid cloud environment.