Read this article in full on the CHI Software blog.
When two groundbreaking innovations merge into one, the world shutters just a bit. An inevitable change happens, and we begin dealing with certain things differently.
Today, we will find out how the duet of AI (Artificial Intelligence) and IoT (Internet of things) impacts people’s lives and businesses. What has changed, and in what way? Our article covers the latest innovations and particular benefits brought by the combo of smart devices and intelligent analytics.
A Match Made in Heaven? Why AI and IoT Work Well Together
Let us start from the very basics to make things as clear as possible. What is IoT in simple terms? This is a system of devices connected via the Internet, acting based on some gathered data. But this data does not work on its own. It should be collected, stored, and processed before being used for decision-making. That is where everything starts falling apart.
IoT is surely (and quickly) spreading around the world, and organizations are facing the dilemma of quality data analytics. Meanwhile, the amount of data is growing every second, and the issue becomes more visible. Let us figure out why this issue is happening.
First, there is a cloud problem. Incoming data volumes are so big that clouds cannot scale proportionally. Second, there is a transportation problem. The capabilities of transferring all the data from an IoT device to the cloud are limited. Even if an organization’s infrastructure is well-developed, growing data volumes put sticks in the wheels.
Now let us turn to the recent IoT innovations. Autonomous cars, for example, require real-time decision-making (similar to human driving) by all means. Otherwise, the technology will not bring the desired results.
The same goes for IoT applications in manufacturing. Imagine unexpected delays right in the middle of the production process. How many consequences, in your opinion, may it have?
These numbers will give you an idea of what is going on in the AIoT (AI + IoT) niche at the moment.
AI Algorithms for IoT: What Are the Best-Case Scenarios?
AI is not that simple. It includes several subsections, such as machine learning, deep learning, computer vision, speech recognition, and others, used separately or in combination with one another. And each of them can be implemented to achieve spotless performance of an IoT system.
Retail: Advanced Analytics and Checkout-Free Shopping
Now retailers can track customer traffic within a store using smart cameras that provide information on the shopper’s behavior and choosing process, as well as allow business owners to optimize the store’s space according to shopping patterns of their target audience.
Moreover, autonomous offline stores have already entered the retail industry. Take the example of Amazon Go opened in the US and the UK. Shoppers in such stores use only a special app to enter and nothing more. This checkout-free experience was named “Just Walk Out”.
IoT sensors and cameras placed all over the store capture the products that customers grab from the shelves and add it to a sales receipt in the Amazon app. The total sum is withdrawn from the customer’s payment card synchronized with the app.
A similar experience is now offered by a Lisbon startup Sensei. Just like Amazon Go, Sensei uses a blend of cameras, sensors, and AI algorithms to provide “grab and go” shopping. Apart from obvious advantages, it appears to be particularly beneficial in the pandemic era because it eliminates any queuing.
Healthcare: Body Trackers
The examples of IoB (Internet of Bodies) devices include well-known fitness trackers, implanted devices, or ingestible sensors that can monitor pretty much any aspect of human health. The IoB data is objective and collected throughout the day, which makes it the most trustworthy source of information about the user’s health condition.
Similar solutions are also implemented in sports or other industries associated with physical efforts. By measuring the state of the worker’s or athlete’s body, managers (or coaches) can adapt workload to each person individually.
Ford has employed the body tracking technology used in sports to keep a closer eye on the workers’ safety and reduce the injury rate during the production process. It is a part of a bigger program initiated in 2003 which eventually has led to 70% decrease of injuries happening at the assembly line.
In combination with AI algorithms, smart devices can “make a decision” to perform a certain action with no human involvement. Raw data collected by the IoT systems is transformed into behavioral patterns that simplify people’s daily routines. Here is a market example to demonstrate this AIoT workflow.
Google has launched the Nest learning thermostat that collects data of how people use the device over time and creates scenarios to set up comfortable temperature in any circumstances. It also helps save energy by adjusting temperature when users are not at home.
Agriculture: Monitoring Systems and Automated Machines
Agriculture is one of the industries where the trace of AIoT is the most considerable. Here are some examples:
- Precision farming systems covering various activities (monitoring soil moisture, managing water consumption, optimizing irrigation, and more) with the use of sensors and autonomous machines;
- Ground- and air-based drones applied to monitor crop health, soil condition, and infestation, as well as spray crops or sow seeds;
- Smart greenhouses based on the work of sensors that together build up an environment to control crops’ growth and condition. Greenhouse sensors can autonomously regulate light, humidity, and temperature in response to collected information;
- Agricultural robots can partially replace routine manual tasks by harvesting crops and then sorting the yield. Robots are trained with AI to monitor crop condition and harvest it at the right moment;
- Solutions to monitor animals’ health and location that, for instance, can help farmers separate unhealthy animals and, thus, prevent a disease from spreading.
This list is not final, and it is regularly updated with new features and applications. But none of it would be possible or efficient enough without AI technologies, particularly, computer vision.
Transportation: Autonomous Cars
IoT-powered cars map up an environment around them using the mentioned sensors located in different parts of the vehicle. For example, radar sensors track the position of other vehicles. Similarly, light detection and ranging sensors help measure distances and identify road edges.
There is hardly anyone who has not heard about Tesla cars with self-driving features. However, it is not a complete autonomy, and the car still needs the driver’s presence. The innovation is still a work in progress, and we are sure that exciting news in this regard is coming soon.
Some examples of smart city features:
- Smart parking to help drivers quickly find parking spots and conduct digital payments for parking services;
- Smart traffic management allowing governments to monitor traffic and optimize traffic lights;
- Smart road lighting that dims in late hours when roads are empty;
- Warning systems that can provide early signals in case of hurricane, earthquake, or flood coming;
- Real-time building monitoring that allows citizens to notify officials when repairs are required.
All these features can be implemented in combination with one another and form large and efficient IoT infrastructures. It is a real chance to make urban areas less stressful and reduce negative climate changes.
Some examples of already existing smart cities are Amsterdam in the Netherlands and Neom in Saudi Arabia. Singapore, Oslo, New York, London, and other cities have also been actively applying smart city instruments.
In our original article, we gathered some good-to-know market stats, listed benefits of AIoT implementation, and provided more explanation on each the mentioned use cases. Follow the link to read more.