In a country as vast as South Africa, where transport by road is the norm, it’s imperative that organisations manage their vehicle fleets effectively.

Today, fleet management goes well beyond the traditional realms of vehicle financing and maintenance to include vehicle telematics (tracking and diagnostics of vehicles), driver management, speed management, fuel management and health and safety management.

To derive real value from these activities, organisations need to be able to monitor events in realtime, analyse them on the fly and, if necessary, initiate an immediate action or reaction.

Among the challenges that fleet managers find notoriously difficult to overcome are driver behaviour and external environmental factors.

Solutions to monitor a vehicle’s speed or route have been available for years, and in-vehicle cameras let managers monitor who is in the driver’s seat too.

But bringing this information – and more – into a centralised system that can analyse it without human intervention and autonomously initiate action has largely been missing.

The new world of Internet of Things (IoT) has opened up new opportunities for fleet management. Onboard sensors and cameras can now measure a wealth of inputs, and relay them via an IoT gateway to a centralised analytics engine.

A young South African startup is making this happen: Raphta is developing IoT systems that collect data from onboard sensors and analyse it using the world’s most advanced artificial intelligence system – IBM’s Watson.

The company has partnered with IBM value-added distributor Axiz to develop and deploy a fleet management system that uses IoT and artificial intelligence to make drivers’ life on the road safer and more productive.

The challenge

Drivers travel long distances over less-than-ideal roads, often in conditions that are not ideal.

Driver fatigue, security issues, plain bad driving, poor road conditions and adverse weather are just some of the factors that can contribute to delays, vehicle damage, illegal actions, security threats or accidents.

Personal safety is a huge issue: if a driver falls asleep at the wheel, or even loses concentration due to fatigue, the consequences could be fatal.

The state of the road, and other environmental conditions like the weather, could also have an effect on driver behaviour. Managers can’t expect drivers to cover the same ground if the road is in poor condition, or they are experiencing adverse weather conditions.

The solution

Fleet management systems seek to help fleet owners to recognise and mitigate the factors that affect driver safety, behaviour and productivity.

Research from Berg Insight indicates that it’s a fast-growing market, set to increase at a compound annual rate of 10,8% to reach 1,5-million units installed in South Africa by 2020.

Now artificial intelligence is set to make fleet management more proactive than ever.

Raphta, supported by Axiz, is breaking new ground in the development of systems that use Internet of Things (IoT) technology coupled with AI that could revolutionise the fleet management market.

Tshidiso Radinne, founder and CEO of Raphta, explains that the company has partnered with a hardware supplier to roll out the solution to a 10 000 vehicle-strong fleet.

The system gathers data from sensors and cameras, feeding it to Raphta’s Shuri platform cloud running Watson on the IBM Cloud.

Watson is IBM’s AI engine, the most advanced AI and machine learning system in the world. Companies can leverage the power of Watson analytics via the cloud, and now a set of APIs (application programming interfaces) allow them to analyse a host of rich new data sources.

Raphta is using Watson Visual Recognition linked to onboard devices and linked via an IoT network to perform a number of sophisticated analyses.

Radinne explains that a dashcam monitors the driver’s face, with analytics determining what state the driver is in.

“For instance, we can tell if the driver is fatigued, and this data can then be used to build a risk analysis pertaining to the driver’s performance,” Radinne says.

This gives fleet managers the opportunity to take the driver’s physical condition into account when planning, or to intervene if a safety risk is flagged.

“The system is also used to analyse environmental and road data,” Radinne adds. “For example, it can pick up potholes or other features of the road’s condition.

“Lane detection can monitor if the driver leaves his lane; and the system can even pick up if the driver is tailgating.”

Raphta developed the system in the IBM Cloud using Watson Studio, with the Watson Visual Recognition API from Axiz and the IBM Bluemix Console.

Its home-grown Shuri platform is the foundation of the analytics system that integrates with Watson Studio to manage workflow and data modelling. Shuri is coupled with Watson Visual Recognition API and Raphta’s own proprietary AI computer vision algorithms

The development of a smart fleet management system, based on visual recognition and incorporating a strong AI focus, is an important one for the South African transportation market. Fleet operators are increasingly shifting their focus from vehicles to drivers, recognising that how well drivers perform behind the wheel can pay dividends in the long-term.

Better monitoring of driver behaviour not only encourages safe driving habits, reducing risks and accident rates, but some insurance providers offer incentives as well.

The first 1 000 vehicles to be deployed will roll out with the embedded sensors this month (September), and the balance will come online over the coming months.

What is IBM Watson?

AI systems are designed to operate in a human context, on human forms of expression, analysing language, movement and even facial expressions.

Watson is the AI system developed by IBM, allowing all organisations access to its sophisticated machine learning abilities to create smarter processes that improve business operations.

Watson is constantly being trained by experts in industry verticals like healthcare, education, finance, transportation and energy so it can be quickly and seamlessly embedded into enterprise workflows.

The days of just structured data being important for enterprise systems are long over, which is why Watson is able to ingest, enrich and normalise a wide variety of data types without additional integration.

Watson on the IBM Cloud

Watson on the IBM Cloud allows developers to integrate the world’s most powerful AI into their applications, to store, train and manage data in the most secure cloud.

This integration allows for the seamless integration of Watson services into digital applications, products and operations, giving organisations access to data-driven insights that improve business outcomes and produce real value.

Watson can ingest, cleanse and comprehend a variety of data types while optimising for the robust AI workloads required to extract meaningful insights.

Watson provides the largest base of industry offerings across the board.

IBM Watson Studio

IBM Watson Studio accelerates the machine and deep learning workflows required to infuse AI into your business to drive innovation.

It provides tools for data scientists, application developers and subject matter experts to collaboratively and easily work with data and use that data to build and train models at scale.

Watson Machine Learning allows for the use of trusted data to put machine learning and deep learning models into production.

Developers are able to leverage an automated, collaborative workflow to grow intelligent business applications easily and with more confidence.

Watson APIs

Watson APIs (application programming interfaces) is IBM’s suite of enterprise-ready AI services, applications and tooling.

With Watson APIs, developers can integrate AI into their most important business processes, informed by IBM’s rich industry expertise.

South African developers can source Watson APIs, along with a wealth of support options, from value-added distributor Axiz.

Watson Visual Recognition

The IBM Watson Visual Recognition service uses deep learning algorithms to identify scenes, objects, and faces in images you upload to the service.

Developers are able to create and train a custom classifier to identify subjects that suit their specific needs.

Who is Raphta?

Raphta is a deep technology company focused on its flagship video intelligence platform Shuri that processes, analyses and makes decision out of video and telematics data collected from vehicles for the fleet management, logistics, auto-insurance and automobile industries.

Raphta is also engaged in fundamental and applied research in the areas of deep learning and quantum machine learning to help corporates solve problems and accelerate their digital transformation in the medical, energy, engineering and finance industries.

The company’s mission is to build radically different and better technologies that can solve complex problems and foster inclusive innovation.

A world-class management, R&D and engineering team comprises top skills from fields like AI, machine learning, computer science, software engineering, quantum and optics engineering.

Founder and CEO Tshidiso Radinne explains that the team may be small – there are 12 people on board at present – but it boasts an extremely high level of skills.

“We’ve got some incredible talent behind the solution,” he says. “In fact, we have probably got the best team in the country when it comes to deep learning and machine learning.

“We are mostly Masters and PhD graduates, and include some international skills as well; with eight out of the 12 people in the company fully trained engineers and scientists.”

Raphta has partnered with leading research labs and schools in South Africa and abroad, and has a strong technical advisory board of leading experts in AI, quantum technologies and digital transformation.

The company has a collaborative partnership with Wits University and currently runs operations out of the Tshimologong Precinct innovation incubator.

“So we are able to collaborate and interact with the AI Lab at Wits (School of Computer Science and Applied Mathematics),” Radinne points out.

However, Raphta skills are responsible for the sophisticated algorithms used to build its Shuri platform, and to leverage the visual data for meaningful analysis.

“We are in startup mode, but we are focusing on technology and systems that are not available anywhere else,” Radinne says. “The company is run by engineers and technologists so we just get stuck into the work.”

The company was named after the first metropolis in Africa, a city in Azania, sometimes referred to as the “Atlantis of Africa”.

Why partner with Axiz?

Axiz is a leading value-added distributor based in South Africa, with operations throughout sub-Saharan Africa.

Together with IBM, Axiz has geared up its operations to help partners and customers to understand the world of cognitive computing, and to build systems that meet real-world business needs.

The 18-strong IBM team at Axiz is geared up to work with business partners, collaborating on the creation of new and innovative AI solutions.

The fully-equipped Axiz demo centre is able to run proof-of-concept projects, while pre- and post-sales experts are on hand to assist customers with planning or deployment issues.

“IBM talks about smarter initiatives, which is something no single partner can deploy,” says IBM business unit manager Bradley McCulloch. “So we are working with business partners to collaborate and create new solutions, and help partners to work with one another.

“This is one of our greatest strengths,” he adds.