The Internet of Things is becoming more and more intelligent. Companies are incorporating artificial intelligence—especially machine learning—in IoT applications and are seeing capacity growth, including improving operational efficiency and helping avoid unplanned downtime.
With the wave of investment, the emergence of a large number of new products, and the continuous increase in enterprise deployment, artificial intelligence is setting off a boom in the Internet of Things (IoT) field. Companies formulating IoT strategies, evaluating potential IoT projects, or seeking to derive more value from existing IoT deployments may need to explore the role of artificial intelligence.
Artificial intelligence is playing an increasingly important role in the application and deployment of the Internet of Things, and this shift is very obvious in the behavior of enterprises in this field. Venture capital in IoT startups using artificial intelligence has risen sharply. In the past two years, the company has acquired dozens of companies working in the intersection of artificial intelligence and the Internet of Things. The major suppliers of IoT platform software are now providing integrated artificial intelligence functions, such as machine learning-based analysis.
Artificial intelligence plays an important role in the Internet of Things because it has the ability to quickly extract insights from data. Machine learning, an artificial intelligence technology, brings the ability to automatically recognize patterns and detect data anomalies, information generated by smart sensors and devices, such as temperature, pressure, humidity, air quality, vibration, and sound. Companies have found that machine learning has significant advantages over traditional business intelligence tools in analyzing IoT data, including the ability to predict operations 20 times in advance, and is more accurate than threshold-based monitoring systems. Other artificial intelligence technologies such as speech recognition and computer vision can help extract insights from data that required human review in the past.
The powerful combination of artificial intelligence and IoT technology is helping companies avoid unplanned downtime, improve operational efficiency, launch new products and services, and strengthen risk management.
In many departments, unplanned downtime due to equipment failure can cause serious losses.
Predictive maintenance-the use of analytical methods to predict equipment failures in advance in order to arrange orderly maintenance procedures-can reduce the economic losses caused by unplanned downtime. In the manufacturing industry, predictive maintenance can reduce the time required for planned maintenance by 20-50%, increase equipment uptime and availability by 10-20%, and reduce overall maintenance costs by 5-10%.
Because artificial intelligence technologies—especially machine learning—can help identify patterns and anomalies and make predictions based on large amounts of data, they have proven to be particularly useful in achieving predictive maintenance.
AI-driven IoT can do more than help avoid unplanned downtime. It can also help improve operational efficiency. Part of the reason is that machine learning can produce fast and accurate predictions and insights, and artificial intelligence technology can automate more and more tasks.
For example, for Hershey, controlling the weight of products in the production process is crucial: every 1% increase in weight accuracy can save more than $500,000 in a batch of 14,000 gallons of products (such as Twizzlers). The company uses the Internet of Things and machine learning to significantly reduce weight changes in the production process. The data is collected and analyzed for the second time, and the weight change can be predicted by the machine learning model, so that 240 process adjustments can be made every day, and before the installation of the ml-driven IoT solution, only 12 process adjustments need to be made every day.
Artificial intelligence-based predictions also helped Google cut 40% of data center cooling costs. The solution is trained based on data provided by sensors in the factory to predict the temperature and pressure in the next hour to guide actions to limit power consumption.
Machine learning generated insights that persuaded a shipping fleet operator to take a counter-intuitive action, saving them a lot of money. The data collected from sensors on the ship is used to determine the correlation between the amount of money used to clean the hull and fuel efficiency. The analysis shows that by cleaning the hull twice a year instead of once every two years-thus quadrupling the cleaning budget-they will ultimately save $400,000 due to improved fuel efficiency.
The combination of Internet of Things technology and artificial intelligence can form the basis for improvement and ultimately form new products and services. For example, in terms of General Electric’s drones and robot-based industrial inspection services, the company hopes that artificial intelligence can automate the navigation of inspection equipment and identify defects based on the data captured by the inspection equipment. This may result in safer, more precise, and save customers up to 25% of inspection costs.
At the same time, Rolls-Royce plans to launch a new product soon, featuring IoT aircraft engine maintenance services. The company plans to use machine learning to help it discover patterns and determine operational insights that will be sold to airlines. Automobile manufacturer Navistar hopes to use machine learning to analyze real-time networked vehicle data, creating new sources of revenue in vehicle health diagnosis and predictive maintenance services. According to data from Navistar’s technology partner Cloudera, these services have helped nearly 300,000 vehicles reduce downtime by 40%.
Many applications that combine the Internet of Things with artificial intelligence are helping organizations better understand and predict various risks, and automate quick responses, enabling them to better manage worker safety, financial losses, and cyber threats.
For example, Fujitsu has tried to use machine learning to analyze data from networked wearable devices to estimate the potential threat heat stress accumulated by its factory workers over a long period of time. Banks in India and North America have begun to evaluate the ability of artificial intelligence to identify suspicious activities in real time through networked surveillance cameras on ATMs.
Auto insurance company Progressive is performing machine learning analysis on data from connected cars to accurately price its usage-based insurance premiums to better manage underwriting risks. The city of Las Vegas has turned to a machine learning solution to ensure the safety of its smart city program, with the goal of automatically detecting and responding to threats in real time.
For cross-industry companies, artificial intelligence may increase the value created by the deployment of the Internet of Things, thereby achieving better products and operations, and gaining a competitive advantage in business performance.
Managers considering new IoT-based projects should be aware that machine learning for predictive capabilities is now integrated with most major level (in other words, general-purpose) and industrial IoT platforms, such as Microsoft Azure IoT, IBM Watson IoT, Amazon AWS IoT, GEPredix and PTCThingWorx.
More and more turnkey, bundled or vertical IoT solutions utilize artificial intelligence technologies such as machine learning. For example, for a connected car use case, BMW’s CarData platform can access data shared by car owners and the AI function of IBM Watson IoT. In the consumer goods and retail industries, many replenishment automation and optimization solutions use machine learning to predict demand and optimize inventory levels. Telematics solution providers in the auto insurance industry are integrating machine learning to create more accurate risk models and predict claims behavior.
Perhaps it is possible to use artificial intelligence technology to get more value from the deployment of the Internet of Things, and the design of the deployment of the Internet of Things does not consider the use of artificial intelligence. For example, a Hungarian oil and gas company applies machine learning to sensor data, which has been collected during diesel production. The analysis allows the company to more accurately predict the sulfur content of the fuel and helps determine process improvements, which can currently save the company more than $600,000 per year. Companies may already be using major levels and industrial IoT platforms are providing new capabilities based on artificial intelligence, which may help increase the value of existing deployments.
In the case of the Internet of Things, machine learning can help companies capture the billions of data points they have and attribute them to meaningful content. The general premise is the same as for retail applications – review and analyze the data you collect to find patterns or similarities that can be learned in order to make better decisions. The Internet of Things will also generate big data, but artificial intelligence is just a technology that makes these big data useful and meaningful for an industry. There are a large number of fields and business niches that can gain the advantages of such coexisting – two technologies. It’s time for the machine to point out where the real opportunity is.