x

SUBSCRIBE TO OUR BLOG

Advice, stories, and expertise about work life today.

x

THANK YOU!

We know your inbox is protected space, so we promise to send only the good stuff, twice a month.

Features Editor's Picks What HERE Does News and Updates

Robotics give HERE data validation the royal treatment

- Update 31/03/2017: We were invited to talk about robotics on BBC World Service earlier today. You can listen to the broadcast here.

 

Robotics have moved from the realms of science fiction into the day-to-day operations of everything from car manufacturing to military services. Now, HERE is using robotics to more efficiently validate its data, ensuring better, more accurate services. And the robot’s name? Princess Estelle.

 

We spoke with Dennis Kemps, Product Validation Manager at HERE, who explains the importance of data validation:

“We are responsible for ensuring that the quality of any HERE product meets customer requirements. We have to make sure that the data is accurate and that the services will work correctly.”

PrincessEstelleGif

The validation process takes a number of factors into account - ensuring the copyrights are correct, that the barcodes are working, that the product functions, and finally, making sure that if you put it in a target system, it works.

He adds, “Think of it this way: if you were selling a brand new car, and a customer came in and saw a dent in the door and a flat tire, it would not only be bad for the customer, but also hugely embarrassing for the seller.”

Testing, testing

Dennis is currently working on a project with a carmaker, delivering the company’s first dynamic map. What, though, does this involve?

Dennis explains: “Our testing begins with an initial integration test, before moving on to end-user testing, where we have to confirm the content and functionality from a customer perspective.”

“Our partners have to be sure that our data works properly in their navigation system, so we started creating tests which were done manually. These tests took in the content itself, the search function (can we actually find the data in the product) and route guidance – so, if I type in a destination, is it accurate, and is it the most efficient route.”

Dennis also stresses that seeing the data written down and appearing on screen are very different. Until you test it in action, for example, there’s no way of telling whether the zoom level is right for the city being navigated.

Human error

According to Dennis, an enormous amount of time and effort is required in order to validate map data manually. He explains:

“Just take one aspect of the map – are all of the countries in Europe available, and can I navigate to and from those countries? This means that if you want to check whether you can navigate to and from each European capital, you are already testing 50 x 50 locations.”

“This is extremely difficult for a human. Still, we want to offer that extra layer of testing, and make sure that from an end-user perspective, all the desired content is there.”

pe2

Dennis explains that while there are ways to test that this data exists on the system, it’s difficult to replicate how an end-user would test and experience these features. This is where Princess Estelle comes in.

Princess Estelle, which consists of a rig, a robotic arm, and an industrial camera which records everything, was designed by Dennis and the team. He explains: “We started with a pilot where she had to enter and verify multiple addresses, which she managed very quickly.”

Now, Princess Estelle can test the database content and the functionality of the system, using it like a customer (only faster). The testing that Princess Estelle does leaves a far smaller margin of error.

Dennis explains: “Imagine if you’re living in Spain and want to navigate to France, only to find that you’re unable to do so using the vehicle’s system. Of course, this would be a real problem, and one that would usually only be identified after an enormous amount of human testing.”

Princess Estelle, however, can quickly and effectively test data points all over Europe and ensure that they are useable for the customer. Dennis adds, “She can calculate 10,000 or so routes in three to four days.”

 

The royal treatment

So, what does Princess Estelle actually do? Dennis explains:

“She can perform destination entry, including special characters. She can reposition herself, allowing her to place herself in any of the locations and test navigation to and from them.”

“Once one part of the validation is complete, she will automatically move onto the next step, making her very efficient for testing.”

pe3Dennis also describes how Princess Estelle can determine whether or not routing is successful, calculating how long a route is and comparing it to previous products and checking for inconsistencies.

If routing is unsuccessful, Princess Estelle records the process so that engineers can review it later on. Princess Estelle’s road recognition also captures traffic data. This is only scratching the surface of the actions the robot can actually carry out, with the team currently working on a ‘pinch’ movement with two ‘fingers’, for zooming.

Dennis adds, “We only started working with Princess Estelle this year, and already the potential is huge. We would like to roll this out to other automotive systems and use her more in the future.”

Looking forward

When asked how the robot has impacted the team, Dennis says, “She allows us to add more depth to our testing. She’s not there to replace people, but to take away the repetition of certain parts of the job and make it easier.”

“She emphasises our effort and willingness to think outside the box in order to provide greater testing and better services. We want to make our client happy, and Princess Estelle helps us to do this.”

Would you be interested in finding out more about Princess Estelle? Let us know in the comments below.

instagram

Urban Mobility

Expand your city network and provide a seamless multi-modal experience on-demand

Download white paper

Sign up for our newsletter