For industrial, automotive and manufacturing robotics applications, 5G technology offers high speed, reliable, dense and low latency connectivity that is suitable for fixed location or mobile use cases.
When discussing the exciting applications for 5G, healthcare, autonomous vehicles, drones and high-bandwidth consumer services get top billing, but 5G is a game-changing technology for machines and robots because it is the first wireless technology with low latency to enable collaboration and high data rates to exchange large amounts of information.
As robots use artificial intelligence (AI) to direct their operations, the fast data rates of 5G will facilitate communication to AI servers for training or inference.
With 5G network slicing, myriad use cases with vastly different performance characteristics can share the same network infrastructure. Slices can be provisioned for applications such as rescue robots, which require high-speed, high-bandwidth performance, or driver-less robotic cars that need low latency, or collaborative manufacturing robots that require massive connectivity. Each can occupy their own network slice, making 5G the one technology that can power innovative robotic applications.
Robotic OEMs have a choice for 5G connectivity. They can build their own chip on board (CoB), assembling all the necessary components, for example, modem, antenna, RF and IF front end, to engineer the interface. The other option is to use a pre-engineered 5G module that is designed for a wide range of applications. Configuring this module for use in an industrial robot provides a low barrier to entry and is a viable choice for robotic OEMs that add value with functions and features to robotic systems.
Either way, these 5G connections need to be tested once built into the robot. Even pre-tested modules, once packaged together with the rest of the components and integrated into the robot, can experience co-existence issues, signal interference or other wireless quality issues. These conditions can affect the overall functionality of the end product.
The testing of 5G is not a new requirement for OEMs, but testing industrial robots may be challenging. One reason for this is that up until now the OEMs and their manufacturing lines were set up to build and test robots with no cellular capability. As a result, incorporating testing of 5G will require an update to the test set up and fixtures.
The 5G technology offers a flexible physical layer design supporting a range of frequencies, diverse bandwidths, sub carrier spacings and modulation schemes. Since 5G testing is new for many robot designs, newer test cases need to be added to verify some newer features within each test area. This could be research and development (R&D), design validation test, sampling, production or service centres.
All of this requires rethinking the traditional test methodologies and making use of cellular signalling or non-signalling testing to ensure quality product and performance testing.
It is important to differentiate between the two cellular testing techniques – signalling and non-signalling.
Non-signalling testing is conducted at nearly every stage of product development, although it is highly targeted and optimised for use at manufacturing. The technique makes use of an RF signal generator and analyser to predominantly focus on calibration and verification of device RF transmitter and receiver performance. This type of testing is performed in non-call processing mode and makes use of chipset-specific test mode to measure pre-defined transmission patterns, thereby minimising test time and reducing the cost of test.
On the other hand, signalling testing is used extensively during product development stages such as R&D and design verification testing and uses a base station emulator to establish an end-to-end user plane call with the device under test (DUT). This approach is extended, as it uses a combination of control plane signalling and user plane traffic to measure the true performance of the DUT under realistic test conditions, but aids in verifying each layer of the protocol stack to perform comprehensive device testing.
Both test methodologies are designed to accomplish distinct test needs and are optimised for use across different product life cycles. However, the choice between the two is largely driven by parameters such as hardware design, test focus, test type, test time, cost of test and beyond.
As 5G continues to be widely adopted, the subsequent section highlights the relevance of 5G testing to ensure the quality of industrial robots.
When building products with the self-engineered CoB design, protocol and functional testing are integral parts of the product design and development stages to ensure the device goes through extensive feature testing and software regression prior to manufacturing.
However, the test dynamics change when building products using third-party cellular module/RF subsystem/antenna modules. This is because, even if the modules sourced are certified, it does not guarantee the final product will function as desired in the field. Once packaged together the components can notably impact each other and the overall behaviour of the product in real- world network configurations.
The table highlights some of the common issues.
Each of the functional issues in the table could significantly decrease the communication reliability, and thus increase latency and affect speed.
This interconnectedness and non-mutual exclusivity between components strongly marks the relevance and need for 5G quality testing in the pre-production and production stages to test the product in its entirety. It is needed to validate:
* Basic device registration and call procedures ensuring end-to-end product functionality, as the hardware components and system firmware could affect the robots’ wireless performance
* Antenna performance, as final product casing, implementation and tuning errors could cause signal loss and affect the antenna radiation pattern, leading to communication failures between the robot and the 5G network
* RF performance, to ensure signal quality is not heavily degraded under realistic channel conditions with acceptable levels of signal transmission and reception as defined by 3GPP and other certification bodies
* Data throughput, as in real-life conditions the device must be able to sustain and handle varied user plane traffic and quality of service requirements
* Real-world user experience features such as mobility and handover
* Advanced functional features such as embedded SIM, multi-SIM testing for immediate connectivity.
Compared to legacy cellular technologies 5G offers a lot more flexibility and an extensive feature set. Thorough product testing can help ensure holistic product quality across different deployment models. Having outlined the importance of quality testing, it becomes equally important to choose a test that not only supports versatile functional tests, but also improves parallel test efficiencies. Consensus among the production test team and the test vendor can help the vendors to develop tools that support quick Tx-Rx (transceiver-receiver) diagnostic tests and support automated testing to reduce time to market and control the cost of testing.
