Software testing for wearable devices is becoming a major challenge

By: Krishna Kishore, Senior Architect and Debabrata Sahoo, Lead Engineer, Sasken Technologies

Wearables have, long, moved from being futuristic to becoming popular due to convenience as they are less intrusive and more like an extension to the human body. They have become common place now to the extent that consumers and professionals have started depending on them for decisions related to diet, nutrition, exercises, health, etc., In addition to this, wearables such as ring/wrist scanners are proving to be essential for professionals working in warehouses and industrial facilities to keep their hands free while moving materials or scanning codes. For professionals working in mines and heavy equipment areas, wearable devices such as smart helmets help workers remain connected event underground and detect hazardous gases or leaks thus ensuring their safety.

Given the critical functions that these wearables have to serve, the hardware and software have to be tested thoroughly. Wearables differ from handsets in display size, shape, expected long battery life, the need for robust algorithms, configurations, and a higher dependency on connectivity technologies. Since the display size is small and of various shapes, testing cannot always be efficient for the small icons in UX design. It is a challenge to validate usability of these devices and provide promising user experience. In wearables such as smart glasses where display is head mounted, testing becomes strenuous for the tester.

Many wearables such as smart glasses, smart bands, smart watches, and smart helmets operate based on voice based commands. Wearables in which voice is the only way of input, it becomes difficult for the tester to differentiate failure due to non-detection/recognition of voice or due to failure in functionality.

We have listed below some use cases of wearable devices along with the unique challenges they harbour:

Detection of Hazardous gases

To test detection of hazardous gases such as CO, CH4, CO2, identifying the different kinds of gases possible in an environment and making them available in the lab to test is a challenge. Moreover, the wearable device has to be tested with all possible hazardous gases because the set of gases may change in each environment. Another challenge is to set the threshold of gas level in air and testing against it.

Collision

Wearables such as mining helmets have to be tested to check if they detect collision correctly. All possible test scenarios have to be identified to test for possible false alarms such as alert based on removal of helmet or hitting a stone or wall accidentally.

Battery

Battery life depends on many factors such as temperature. Since wearables are used extensively in outdoor places and are exposed to extreme weather conditions, battery life has to be tested in all possible simulated scenarios which are not practical.

Algorithms

Algorithms are most critical in wearables. They have to be tested thoroughly considering all possible datasets. While functionality with each dataset is important, performance for all datasets need to be consistently optimal.

Since algorithms work based on values from sensors such as accelerometer, gyrometer, magnetometer, etc. and behavior of sensors depends on the environmental conditions, there is a need to identify all possible conditions. The behavior of one sensor may differ when it is used individually from the scenario to when it is used along with values from other sensors in case of sensor fusion. The parameters of each sensor such as sensitivity, supported range, etc. differ even with same sensor type from different vendors. Test case design should consider this aspect.

Gestures such as wrist and motion are being used in wearables therefore defining accuracy of gesture algorithms is a challenge because pattern of various users will differ.

Non-standard UI design

Because of non-standard UI design, test approach on each wearable may differ. Hence, effort of testing is high due to minimal reuse.

Certifications

Since wearables are used close or in contact with body, for safety purposes regulatory authorities define thermal, radiation, and electrical conditions. For critical segments such as medical, the connectivity and functionality has to be reliable. To take care of these aspects, certification bodies mandate wearables to go through certification processes before getting into the market. Certification in various geographies needs to be taken care of.