Robotic process automation technology helps test instruments improve efficiency

By Andrew Herrera, RF Test Software Product Marketing Manager, Keysight Technologies

Product innovation is driving new features and capabilities, driving up research and development (R&D) costs. Electronics companies spend a lot of money on components, labor, and testing during the R&D process. Testing covers every stage of the R&D process. Before a new design element is incorporated into a product, manufacturers must test the performance of that element. Even if they only replace a small component with a lower-cost or higher-performance component, they must retest it.

This requirement applies not only to components but also to software. Since many products are controlled by software, the output indicators allowed by the design measured by the software determine the performance of the product. For wireless solutions, this may be an extra step in the development process. To ensure that they do not exceed the limits set by the wireless standards, engineers usually need to perform many repeated tests.

Design, testing, and verification are not only time-consuming, but also require repetitive processes, resulting in high costs. Design verification engineers must ensure that the solutions they design can perform well in harsh environments while minimizing the costs of the manufacturing and verification stages. Figure 1 shows a development process for the product life cycle. Each step requires testing, debugging, and verification.

Figure 1. Product life cycle development example

Many applications use automation to improve efficiency. Robotic process automation (RPA) helps reduce repetitive operations and manual labor in the testing process, such as clicking on software and changing test instruments. RPA allows engineers to have sufficient energy to handle other projects or tasks during repetitive testing, thereby speeding up hardware verification. Before evaluating whether RPA can save time for design verification engineers, we must first understand testing and the associated costs.

Time and cost of hardware validation testing

When it comes to the time and cost of test design, there are many repetitive tasks to consider:

• Test hardware under conditions that simulate real-world conditions

• Validate hardware to ensure it meets technical specifications, user expectations, and local environmental regulations

• Debug hardware to ensure it performs as expected under normal and abnormal conditions

• Ensure that appropriate safety measures are in place and that they comply with appropriate safety standards

The time and cost of each test depends on the project or task at hand. Let’s take the example of three hours per test. The four examples above would require an engineer to spend 12 hours on four tests. This assumes that the engineer performed all steps correctly, each measurement yielded the expected results, and no errors were made during instrument changes and adjustments. If not, the test time could extend from three hours to four or five hours. If the engineer’s repetitive processes can be automated, a lot of time can be saved.

Verification engineers perform a measurement in a configured environment, then stop to set up the next environment and perform the same test again. They repeat this process until all scenarios are tested. Figure 2 is an example of a workflow for hardware verification on a device under test. Throughout the process, engineers need to switch test instruments, change probes and hardware, and adjust settings. Not every test uses the same instrument or software, and changing settings can cause testing time to increase.

Figure 2. Example workflow for hardware verification on a device under test.

Test environments vary, and engineers have a variety of manufacturing test instruments on their test benches. If the test environment does not support certain software or instruments, testing may have problems. In this case, the instrument needs to be switched or a new compatible instrument must be ordered, wasting more time and cost, resulting in more test time required for each task.

Scaling each configuration to 10 or 20 tasks can cause project delays and increased costs. As we go deeper into each task, we see the true cost of developing and validating tests during the R&D process, and we also see a common cost issue in each test: time.

Save time on hardware validation testing

Every business owner knows that if a project or task takes a lot of time, it means more costs. However, rushing engineering hardware verification or cutting corners can lead to catastrophic failure. Therefore, the best way to shorten the time spent on each task is to complete the task more efficiently. High-volume test environments need to be more efficient. Measurement steps must focus on execution speed. To optimize testing, developers need to be in control. Test engineers must be familiar with the following optimization obstacles:

• Instrumentation solutions may come from multiple manufacturers

• Pre-configured analysis routines often lack speed and flexibility

• The test application performs more measurements than necessary

• Equipment with parallel analysis capabilities may be complex and inefficient to operate

Every step in testing carries the risk of human error. For these optimization barriers, instrument changes or measurements are at risk of error if inputs are incorrect. Performing a large number of tedious, repetitive tasks can cause engineers to overlook small steps. These seemingly minor issues can introduce significant errors into test validation, leading to a huge risk of human error.

Once such an error occurs, engineers need to spend more time retesting, which means more cost and time invested in R&D. RPA helps speed up the debugging and validation process while minimizing human error. Automation eliminates the risk of small adjustments being overlooked because it strictly executes every step, no matter how tedious it is.

We evaluate RPA in a test instrument environment. RPA must be able to configure and build complex workflows through parameter configuration. If automation requires an engineer to enable changes, its efficiency cannot be guaranteed. With this in mind, software automation needs to enable remote access so that someone can check the test status and enable parameter changes if the engineer is not nearby. Using test instruments from multiple manufacturers will lead to additional automation challenges and optimization barriers.

Finally, all of these requirements are challenging in themselves, and if engineers have to write code to enable this process, it will add more cost and time, leading to inefficiencies. Therefore, automation software should allow for recording, playback, and sharing of automated tests, all without having to manually write software code.

In a test, engineers need to spend several hours to perform different measurements. RPA can help engineers complete repetitive tests by changing instrument measurement parameters or switching different application software to perform different measurements. This will improve performance and efficiency, allowing test engineers to focus on the next project. Because the test time was halved, the project was completed on time or even ahead of schedule, so that only four tasks could be completed in 12 hours before, and now eight tasks can be completed. In the manufacturing process, better use of test time can achieve a higher return on investment.

RPA is ideal for testing instruments

In addition to improving the efficiency of testing tasks, automation also helps improve employee performance. Spending long hours on repetitive tasks can cause stress to employees and affect their performance. Since automation enables better performance and results, engineers can work on multiple projects without worrying about delays or missing performance targets.

Integrating RPA into the test instrument environment helps increase innovation through hardware verification and testing, and promotes innovation efficiency through improved employee satisfaction. The purpose of RPA is not to increase the difficulty of operating test instruments, but to accept and introduce repetitive test tasks. RPA can help us save valuable time.