by Bernard Marr
In industrial applications, more data makes it more likely that machines will make accurate predictions – by “crowd sourcing” from its own machinery, KONE ensures that its systems are trained using the best possible real world datasets.
KONE has leveraged one of the key opportunities offered by the “data age” by becoming a data provider. It is effectively monetizing its own data by packaging and selling it to other organizations. It recognizes that its own data is valuable due to its power to drive change and efficiency.
Sources
Forbes, Internet Of Things And Machine Learning: Ever Wondered What Machines Are Saying To Each Other?: https://www.forbes.com/ sites/bernardmarr/2017/02/21/how-ai-and-real-time-machine-data-helps-kone-move-millions-of-people-a-day/#5a69c1365f97
IBM, More than 1 million connected: https://www.ibm.com/watson/ stories/kone/
QZ.com, Listen to internet-connected elevators talk about how their day’s going: https://qz.com/910593/listen-to-internet-connected- elevators-talk-about-how-their-days-going/
Smart group control systems – AI work in elevators starting in late 80s: https://www.bernardmarr.com/default.asp?contentID=694
45
Daimler AG: From Luxury Personal Cars To Passenger Drones
Daimler AG, the German parent company of the Mercedes-Benz as well as Smart car brands, has a long history of building luxury and consumer vehicles, trucks and buses since its predecessor companies merged into Daimler-Benz AG in 1926.
Today, as well as being renowned for its precision-engineered automobiles, it is investing heavily in automation and fourth industrial revolution technology, from its design and production operations to the vehicles themselves.
Machine learning is playing an integral part in every part of this transformation, helping to streamline processes, cut down on waste and remove human error from many equations.
What Problems Is Artificial Intelligence Helping To Solve?
Mercedes-Benz is using artificial intelligence (AI) to create efficiencies in vehicle production, transportation and passenger transport.
Vehicle design and manufacture is a labor intensive and costly process, involving high-tech plants and equipment and large workforces.
Equipment breakdown and human error can lead to wasted resources, costly delays and injury.
In addition, the changing ways that we are using personal transport is leading to problems for car manufacturers that are sticking to traditional models. Particularly in cities and urban areas, a move away from car ownership and towards ride sharing and public transport means a declining customer base. Often this is driven by an increased awareness of environmental issues, as well as changing urban landscapes that are becoming less friendly to owners and operators of personal vehicles.
How Is Artificial Intelligence Used In Practice?
Daimler Trucks has unveiled the Future Truck 2025 – which it claims is the world's first self-driving heavy goods vehicle. Although it still contains a cabin to transport a crew (who may be required for loading and unloading of cargo) it can navigate fully autonomously, which the company says will lead to improved road safety and lower fuel costs.
As far as personal vehicles go, Mercedes-Benz is also investing in in-car AI, currently known as MBUX (Mercedes Benz User Experience) to free up drivers from repetitive or distracting tasks while on the road. MBUX can carry out tasks such as predicting a likely destination and automatically engaging navigation systems, and even activate climate controls via indirect commands such as detecting when the driver says “It's hot.”
It has also announced its Luxury in Motion car as the future of autonomous driving. Its elegant and spacious interior is designed to feel more like a lounge than the cabin of a vehicle, allowing its executive and VIP passengers to spend their journey productively and arrive at their destination fully refreshed. The concept is to provide a “mobile living space” and reimagine the concept of personal transport.
This focus on automation extends to the design, production and sales of vehicles. Cameras, sensors and Internet of Things technology give the business a real-time overview of its stock and the operating efficiency of its machinery. It means that each vehicle can be built to a customer's specific requirements, while remaining in a mass-production environment. This allows it to offer a feature through its Mercedes-Me app, which it calls Joyful Anticipation – allowing buyers to track the progress of their in-production car as it passes through the assembly line.
On the sales side of the business, Daimler also lets potential customers who spot their dream car on the streets to snap a picture using their Car Detection App. The image will then be analyzed using image recognition algorithms, and the potential customer will be informed of the exact make, model and specifications – as well as where it can be bought in the local area.
Daimler's ideas about how transport will change includes an understanding that we are likely to continue to move away from the concept of individual vehicle ownership in the future. No doubt this thinking is behind its purchase of 60% of the MyTaxi ride-hailing service, as well as the Athlon car leasing business.
Looking to the skies, and further into the future, Daimler has also staked its claim to a share of Dubai's plan to offer the world's first drone taxi service. The company has invested £25 million in Volocopter – the German firm that took Crown Prince Sheikh Handan Bin Mohammed on a five-minute maiden flight above the desert in 2017. Autonomous drones will use machine learning to safely navigate past other air-borne objects and react to changeable weather conditions in-flight.
What Were The Results?
Most of Daimler's AI projects are in pilot or prototype stage and there is little data on how effective they have proven to be so far. However, the company-wide focus on intelligent, self-learning technology shows that Daimler is firmly committed to an autonomous future. Self-driving trucks such as the Future Truck 2025 have the potential to improve safety on the roads, while more high-flying initiatives such as Volocopter could help to relieve congestion on the ground in our gridlocked cities.
What Technology, Tools And Data Were Used?
Daimler partnered with Nvidia to design deep learning-based systems three years ago, and the technology developed here forms the basis of its autonomous driving and AI assistant systems.
Data is gathered from the road itself using sensors attached to the vehicle, and is processed through computer vision systems. It is augmented with external data such as GPS and meteorological information.
In its production environments, data is gathered from cameras and sensors fitted to machinery, as well as data from computerized stock control systems, machine data and customer service feedback. 3D printing and virtual reality are also used for design and prototyping.
The Car Detection App uses SAP's Leonardo machine learning platform to analyze pictures of Mercedes cars and tell the sender its make and model, as well as where it can be bought locally.
Key Challenges, Learning Points And Takeaways
As with other auto manufacturers, and leading businesses in many other industries, Mercedes is moving away from its traditional background as a car maker and positioning itself as a data-driven technology company. In the near future its competitors are just as likely to be Google and Apple as BMW or Toyota.
Businesses that lead the way in AI and automation aren't limiting themselves to individual use cases. Technology can be implemented right through a company, from design and prototyping to sales and servicing.
Auto manufacturers are moving away from identical production lines and building the capability to produce bespoke products with the same efficiency as they can with mass-production. AI enables them to cope with the inherent logistical challenges.
Automobile production lines of the near future will be safer, faster and more efficient, thanks to an improved ability to collect and analyze data at every step of the process.
Sources
Business Insider, MBUX: http://uk.businessinsider.com/merced
es- building-its-own-ai-powered-voice-assistant-for-the-car-2018–1?r= US&IR=T
Daimler, AI in car production and manufacturing: https://www. daimler.com/innovation/case/connectivity/industry-4–0.html
SAP, Car spotting app using SAP tech: https://news.sap.com/2018/06/ machine-learning-makes-mercedes-benz-dream-car-a-reality/
46
NASA: Using Artificial Intelligence To Explore Space And Distant Worlds
NASA will launch its next mission to Mars in 2020. It has landed four Mars rover craft on the surface of the red planet so far, starting with the first successful landing of Sojourner in 1997. The most recent landing was by the rover Curiosity in 2011. As artificial intelligence (AI) technology is far more advanced than at the time the last rover was launched, the as-yet unnamed Mars 2020 rover craft will be the most automated and intelligent yet. Its primary goal will be searching for signs that the red planet may once have been home to life.
Beyond this, NASA's deep space probes – such as the New Horizons mission to Pluto and the Voyager missions to the outer reaches of the solar system – have travelled further than any other man-made object from Earth, and continue to send back data, adding to our understanding of the universe we live in.
What Problem Is Artificial Intelligence Helping To Solve?
One of the biggest obstacles in space exploration is the limited amount of bandwidth available for sending information back to Earth. Due to the distances involved, even today these data volumes are measured in mere megabits.
Particularly when exploring the far reaches of the solar system, unmanned spacecraft can often be out of contact with humans for long periods of time. Their ability to make autonomous decisions about what information is valuable to their Earth-bound operators is vital.
Another problem is the limited amount of power available to operate the spacecraft. As they are often far from recharging stations, and even further from the sun's source of solar energy, power usage must be carefully predicted and monitored. Running out of energy on the surface of a distant planet or in the far reaches of interplanetary space means a multi-billion dollar spacecraft becomes a non-responsive and inoperative hunk of metal, plastic and circuitry.
Additionally, when it comes to manned space exploration, problems arise because working conditions in space often put stresses on the human body far beyond those that human bodies are used to coping with.
How Is Artificial Intelligence Used In Practice?
Spacecraft – from deep space probes to planetary landers like the rovers – are equipped with a large number of sensors to capture every possible piece of information about the environments. This isn't because most of that information is useful – in fact, it generally isn't. The vast majority of space is an empty vacuum and the vast majority of planetary surfaces are comprised of lifeless, inert matter, no different from that found on Earth.
Instead, what it is used for is to build an understanding of what is normal, so that interesting, unusual and valuable information stands out. Teaching space-faring machines to recognize this anomalous data is the main purpose of AI work carried out by NASA.
As Kiri Wagstaff, principal data scientist with NASA's Jet Propulsion Laboratory machine learning group, says: “We don't want to miss something just because we didn't know to look for it.
“We want the spacecraft to know what we expect to see and recognize when it observes something different. If you know a lot in advance you can build a model of normality – of what the robots should expect to see. For new environments we want to let the spacecraft build a model of normality based on its own observations, that way it can recognize surprises we haven't anticipated.”
Smart systems also carefully monitor the power usage of spacecraft – particularly the Mars rovers – to determine which systems are using the most energy, and what can be shut down at any given time to ease the burden on the radioisotope thermoelectric and solar generators. Data on energy use can be correlated in real time with the craft's “plans” for what it has to do over a particular time period – travelling or taking readings – to ensure that the 100 watts of power available at any time are used efficiently.
AI-driven robots are also increasingly being used to augment the abilities of human astronauts working in space. Since the 1970s, NASA has been developing humanoid robots that can carry out manual work or provide assistance alongside human crews. NASA currently uses a robotic system known as Robonaut 2 to assist humans carrying out complex technical operations in the hazardous environments of outer space. Robonaut 2 is a modular, humanoid robot equipped with AI-driven image recognition technology.
What Were The Results?
Previous rover missions, which did not rely on autonomous decision making by onboard systems, were constrained by the fact that it took 24 minutes for information gathered by their sensors to reach Earth, and another 24 minutes for instructions based on that information to be returned to the red planet. For deep space probes, that delay is obviously far longer. Thanks to the implementation of AI systems, information can now be acted on almost instantaneously by the rover, meaning it can make up its own mind about which locations are worth investigating. Given the huge cost of operating interplanetary vehicles and their Earth-bound operation centers, this means more productive missions and a greater human understanding of what lies on the “final frontier”.
The smart data-driven analytics engines on board the Curiosity rover were instrumental in helping NASA establish that Mars once was a habitable environment for life. The next rover – scheduled to launch in 2020 – will be built around this technology from the ground up, and its mission will be to find out whether life actually existed on Mars.
What Technology, Tools And Data Were Used?
To sift through the vast amounts of data collected by lander craft and probes, NASA relies on similar tools to those used by today's data-driven online services such as Netflix and Amazon.
Elasticsearch – the open source search and analytics engine – forms the backbone of several AI systems, including those used on the rover but also to capture high-resolution data on soil moisture across wide geographical areas back home on Earth.
It also uses a software system called AEGIS (Autonomous Exploration for Gathering Advanced Science) to determine interesting features such as anomalous rocks that can be vaporized by Curiosity's lasers so that their composition can be determined.
Robonaut 2 was developed in cooperation with General Motors and has almost human levels of manual dexterity. It became the first humanoid robot in space when it was despatched to the International Space Station in 2011. Since then it has received continual upgrades and is currently capable of carrying out many manual, repetitive and dangerous tasks. In the future, it is planned that it could lead the way on missions to other planets such as Mars, with the task of preparing environments suitable for arriving humans. The technology is also available for licensing by other companies, and NASA points to its suitability in a wide range of logistics, manufacturing, industrial and medical roles.
Key Challenges, Learning Points And Takeaways
NASA is pioneering AI to help solve problems in outer space as well as back home on Earth.
Space exploration generates huge volumes of data, and it is far more efficient to use autonomous machines to work out what is worth sending home and what can be discarded.
Technology developed for space exploration often has utilities back home on Earth – and licensing them can help fund the high cost of development and deployment in space.
Sources
NASA, A.I. Will Prepare Robots for the Unknown: https://mars.nasa. gov/news/2884/ai-will-prepare-robots-for-the-unknown/
NASA, Towards Autonomous Operation of Robonaut 2, Julia M. Badger, Stephen W. Hart and J.D. Yamokoski: https://ntrs.nasa. gov/archive/nasa/casi.ntrs.nasa.gov/20110024047.pdf
NASA, Robonaut 2 Technology Suite Offers Opportunities in Vast Range of Industries: https://robonaut.jsc.nasa.gov/R2/
47
/>
Shell: Using Artificial Intelligence To Tackle The Energy Transition
Royal Dutch Shell started out as a shop selling seashells and is, as of 2018, the world's fifth largest company by revenue.1 Its activities span the whole fuel supply chain from exploration and drilling to refining and retail. Shell is a world leader of end-to-end fuel production – exploration, mining, refinery and retail – including oil, gas, biofuel, wind and solar.
A major challenge it is facing right now is energy transition – the move away from fossil fuels towards cleaner sources of energy. However, it maintains that renewables cannot yet supply all of the energy needed by us to function at our current level of activity and comfort.
Striving to deploy artificial intelligence (AI) throughout its business means covering both of these positions. Therefore, Shell investigates smart technological solutions to both fossil fuel mining to drive efficiencies, which in itself cuts emissions, and to its objectives in renewables.
What Problem Is Artificial Intelligence Helping To Solve?
Experts agree that the future of driving is likely to be electric, and reconsidering our relationship with the internal combustion engine has a big part to play in hitting climate change targets.
But one of the reasons most commonly given by drivers for delaying the change is due to a lack of roadside charging facilities.2
Shell is working to increase its RechargePlus charging stations but the “rush-hour” pattern of driving behavior, particularly in cities, is problematic.