AI and IoT: The Building Blocks of Game-Changing Robotics
24 January 2023
The recent press around ChatGPT shows that as AI-powered applications evolve, the next logical step is for software to become more connected and informed by the physical world around it.
I recently caught up with a friend who led a major AI project to develop a kitchen application that could help with recipe recommendations and taste pairings. Their objective was to develop an AI Chef, i.e., a smart application that could recommend new foods to eat or cook based on what was already available in someone’s kitchen.
The project garnered significant interest from the major appliance manufacturers, who saw it as the perfect solution for a “smart” refrigerator or oven. Imagine if your fridge or cooker could recommend things you could make, including new and unusual foods or flavors, based on what was already in your cupboards.
This technology could be used to improve the nutritional value of what you eat, as well as help minimize food waste and the associated carbon impact. It would be a huge win for the consumer, the manufacturer, and the planet.
At first glance, the solution was simple: equip a refrigerator with machine vision, identify items in the fridge, and have the AI Chef make the appropriate recommendations. However, there was a problem. Visually identifying the contents of the fridge didn’t provide all the information needed. For example, while the vision system could identify a 500g tub of margarine, it wouldn’t know how much margarine was actually in the tub: 10g… 200g… 500g…? No matter how much data was inferred from previous usage, an accurate quantity could never be known with only a visual inspection of the packaging.
I recently caught up with a friend who led a major AI project to develop a kitchen application that could help with recipe recommendations and taste pairings. Their objective was to develop an AI Chef, i.e., a smart application that could recommend new foods to eat or cook based on what was already available in someone’s kitchen.
The project garnered significant interest from the major appliance manufacturers, who saw it as the perfect solution for a “smart” refrigerator or oven. Imagine if your fridge or cooker could recommend things you could make, including new and unusual foods or flavors, based on what was already in your cupboards.
This technology could be used to improve the nutritional value of what you eat, as well as help minimize food waste and the associated carbon impact. It would be a huge win for the consumer, the manufacturer, and the planet.
At first glance, the solution was simple: equip a refrigerator with machine vision, identify items in the fridge, and have the AI Chef make the appropriate recommendations. However, there was a problem. Visually identifying the contents of the fridge didn’t provide all the information needed. For example, while the vision system could identify a 500g tub of margarine, it wouldn’t know how much margarine was actually in the tub: 10g… 200g… 500g…? No matter how much data was inferred from previous usage, an accurate quantity could never be known with only a visual inspection of the packaging.
A New Wave of “Old” Technology: IoT
The solution to this problem? “Smart” applications that have more detailed information about physical objects – which is where IoT (aka the Internet of Things) will finally come into its own.
IoT has been around for more than 20 years. It envisions a network of computing units that can be embedded with sensors, software, and other technologies that connect and exchange data with other devices and systems over the internet.
However, it is only in the last few years that we’ve begun to see what is really possible with IoT, pervasive connectivity, and the emergence of truly useful AI.
The long-term benefit of IoT is its ability to enable software applications to measure, sense, and understand the physical world and, subsequently, allow the development of smarter, more efficient physical systems. For example, an AI Chef that will be able to recommend what you can really make from the actual contents of your fridge - not just identifying the box or container, but weighing it in real-time and accurately suggesting possible meals.
As smart systems establish this level of physical interaction with the real world, it will take us into an entirely new realm of hyper-viable robotics and smart systems – one where we can begin to move and manipulate items within the real world more effectively. The interplay of these systems will help us manage and optimize our tasks and environments, impacting everything from food preparation to office work to waste management and even elder care.
Measurement and identification of these real-world parameters drive the kitchens that are at the heart of all Karakuri products. From temperature to mass, time, force, speed, and even viscosity - every part of the physical world around our robotics is measured, logged, and tracked using the IoT technologies we’ve built and developed in-house.
IoT has been around for more than 20 years. It envisions a network of computing units that can be embedded with sensors, software, and other technologies that connect and exchange data with other devices and systems over the internet.
However, it is only in the last few years that we’ve begun to see what is really possible with IoT, pervasive connectivity, and the emergence of truly useful AI.
The long-term benefit of IoT is its ability to enable software applications to measure, sense, and understand the physical world and, subsequently, allow the development of smarter, more efficient physical systems. For example, an AI Chef that will be able to recommend what you can really make from the actual contents of your fridge - not just identifying the box or container, but weighing it in real-time and accurately suggesting possible meals.
As smart systems establish this level of physical interaction with the real world, it will take us into an entirely new realm of hyper-viable robotics and smart systems – one where we can begin to move and manipulate items within the real world more effectively. The interplay of these systems will help us manage and optimize our tasks and environments, impacting everything from food preparation to office work to waste management and even elder care.
Measurement and identification of these real-world parameters drive the kitchens that are at the heart of all Karakuri products. From temperature to mass, time, force, speed, and even viscosity - every part of the physical world around our robotics is measured, logged, and tracked using the IoT technologies we’ve built and developed in-house.
We believe data is the future of the automated kitchen.
This is why we’ve taken things one step further with our Karakuri: Robot Ready partnership program. The program enables seamless integration of all aspects of an automated kitchen, from our /FRYR automated fry lines to the EPOS, KDS, and KMS software platforms that drive the ordering, cooking, and delivery process. Ultimately, Karakuri: Robot Ready will bring together the fully automated kitchen ecosystem, improving the quality, efficiency, and choice of future kitchens.
