Smart cities and the need for testing speed

Antony Edwards, CTO, Testplant discusses how intelligent process automation can make smart cities a reality – before it’s too late

At the moment, any technology that interacts with the physical world, including today’s ‘smart city’ technology, can take anywhere between 10 and 15 years, on average, to comply with regulation. Perhaps even more significant is that companies themselves need to feel confident that the technology will not fail when it is implemented into their infrastructure so that their users can accept technology.

The main reason for this is that in many companies, both current testing practices and approaches simply aren’t up to the job. Most of the methods currently being used are out-of-date, and came into existence solely to provide vague assurances about non-critical enterprise software. The old approaches are still in use because currently there aren’t many viable alternatives that will be approved by both regulation testers and users with enough time to establish a smart city by 2030.

This delay certainly doesn’t align with the world we live in— today’s consumers expect an Uberised-level of service, and to adapt to consumer demand, cities need intelligent process automation now.


Intelligent process automation: the key to making smart cities a reality?

Intelligent process automation is the application of advanced analytics technology, such as artificial intelligence and machine learning designed to automate the testing and analysis of applications. These technologies are enhanced with benefits such as computer vision and cognitive automation that serve to boost their IoT processing capacity and machine learning capabilities, which speeds up two key areas: testing and acceptance.

With intelligent process automation applied to testing, each and every testing hurdle can be addressed significantly faster than is the case with existing test automation. This is because existing test automation currently focuses on automating the test execution, using test scripts that are created manually.

When pre-employed technologies are tested in digital app format, workflows can be exercised while intelligent analytics are collected and used to improve the product at the same time. This is due to deep learning and cognitive technology that significantly boosts efficiency by removing the traditional routine operations of lab testers improving each product by hand.

AI-driven bug-hunting algorithms to identify root problems that the user faces during their own personal experience, and, utilising these algorithms, predictions of future problems can be calculated and fixed in no time. The customer experience is made far more direct when interactions with the product are both simplified and speedier.

A case where AI algorithms need to be thoroughly tested is that of autonomous smart city taxis. When testing autonomous vehicles, these algorithms will auto-generate previously resourced intensive scripts, which will then improve the test script for the car’s AI-powered behaviour after ridding the car’s system of any predicted hazards. The fact that these self-driving taxis are due to soon appear in controlled California environments and neighbourhoods is testimony to how rapidly this technology identifies experience-based faults and corrects systematic errors.

Only software that provides a speedier testing experience for the everyday user will provide widespread and accurate feedback to the supplier. This will ensure that the tested technology is improved at a faster pace, and will be accepted at a wider societal level within what is more likely to be three years, as opposed to ten. User experiences and opinions need to be diversified, and not limited to the ideas of a few select lab-testers that do not represent the diverse population of 2030’s smart cities.


The approval of the public, not lab testers

Without adequate means of both testing new technology and deploying it faster, our cities won’t become smarter – they’ll just continue as they are now, with what is currently new technology quickly becoming out-of-date, and in urgent need of updating.

This inherently fails to address the situations that cities of tomorrow need to be conscious of: rapidly expanding populations, greater life expectancies, and increasingly threatening air pollution. This much-needed technological efficiency will be a challenge to accomplish, given that the public, not just the lab testers, needs to accept new technology prior to it becoming one with the infrastructure. As such, the only logical process is one that connects the smart technology directly with the consumer. Once their experiences are registered, machine learning will quickly adapt smart products to tomorrow’s needs.

But eventually, people will accept and expect smart cities. If innovative automation technology could be integrated within three years, we’d have another 10 more years of technological revolution that could still make it into the smart city of 2030. In this case, the possibilities would be endless.

Related reading