The Internet of Things (IoT) and robotics are often talked about interchangeably. But their differences run far deeper than their similarities.
It’s true that both of these innovations have had a profound impact in the last decade. They’re also highly complementary forms of technology. Around the world, businesses are being transformed by the combination of robots performing physical tasks and IoT sensors capturing data from equipment, processes and environments in real-time.
However, when implementing IoT and robotics in a business, it’s a mistake to view them as a package deal. These two kinds of technology have vastly different requirements when it comes to your underlying IT infrastructure. There are many IoT systems out there to choose from, especially compared to purpose-built robotics platforms - but the former won’t cut it for your robots’ complex needs.
Here’s why IoT platforms aren’t fit for operating autonomous robots effectively.
1: The volume and frequency of data
IoT platforms are typically built to transmit small amounts of data from a large number of devices. For example, you might have thousands of thermal sensors installed on your factory floor that are connected to a single IoT platform.
These sensors often don’t need to transmit data in real-time. You might have thermal sensors on refrigerators or air-con units providing an update every five seconds. This is a sufficient frequency to assure operators that everything is operating at a safe temperature.
In comparison, an autonomous robot is more like a computer server with arms, wheels or legs. There might be hundreds or even thousands of sensors attached to a single robot. It’s essentially the opposite of the IoT scenario - you’re operating far fewer machines, but each one is collecting much greater volumes of data.
With autonomous robots, it’s also more likely this data needs to be transmitted in real-time over low-latency networks. Some of today’s autonomous robots are so sophisticated you won’t even be able to perform basic functions without 5G. Whereas many IoT devices are simple enough to operate on 4G, 3G - or even low power or low bandwidth networks such as LoRa or the Iridium network for satellites.
2: The type of data (and distance travelled)
Robots don’t just transmit high volumes of data at frequent intervals, they transmit very bandwidth-consuming types of data. For example, an operator may wish to remotely teleoperate a robot. To do this effectively, it helps if the operator can ‘see what the robot sees’ using live video streams, LiDAR or point cloud data.
And this bandwidth-consuming information may need to be transmitted over a long distance if the robot has to work in remote or inaccessible locations. This means you encounter the speed of light problem - you’re adding between 20 and 200 milliseconds automatically just because you’re operating the robot from a distance.
Delays in data transmission are frustrating for robot teleoperators. At worst, delays can be dangerous, such as when autonomous robots are used to perform critical tasks in a factory or hazardous environment. Operators may need to quickly seize control of the robots and intervene in an emergency - in these situations, every millisecond counts.
And of course, this data isn’t just streaming in one direction - it’s bidirectional. During teleoperation, control messages also need to be sent back to the robot in as few milliseconds as possible. However, with IoT, a simple instruction like ‘turn off’ or ‘turn on’ can take anywhere from a second to ten seconds and this usually won’t have a major impact on operations.
3: The dynamic nature of robots
A utilities engineer might install IoT sensors on underground smart water metres and leave them for 10 or even 20 years, transmitting critical data without them ever needing to be touched. This means operating instructions can be hard-coded into the IoT devices.
But robots are dynamic machines. They’re not designed for you to ‘set and forget’. The type of data robots need to collect will vary according to circumstances. For example, if you’re trying to remotely teleoperate a robot you’ll need high levels of telemetry streaming to a visual interface - consuming a huge amount of bandwidth.
However, once your teleoperation session ends you no longer want that information transmitting. It should be switched off to save power and bandwidth until another teleoperation session is required. This means you can’t simply code requirements into robots and leave them alone - you need to control their actions and the flow of data dynamically.
The best way of achieving this is through a cloud-based platform like Rocos. Rocos allows you to configure telemetry streaming to suit your needs. If your robot is connected to our platform and you open a browser to request certain aspects of telemetry, that data will start streaming in real-time. But as soon as you close your browser, the streaming stops - minimizing the consumption of bandwidth when idle.
In addition to live telemetry, which is only streamed as required, you can set up storage streams in Rocos - these enable you to capture specific kinds of data continuously for as long as you like. This means that if an incident happens, you can go back to that moment in time and dig into the surrounding data for context. And if the robot goes offline, the data is retained and uploaded to the platform when connectivity is reestablished.
All of this enables more dynamic control that is customized to your business needs - and this could not be achieved with an IoT platform. For secure, real-time, bidirectional data streaming, especially over long distances, a purpose-built robot operations platform is essential.