Microsoft Research: Beyond Kinect, And Open Source Robots

Scanners, sensors and a football-playing chair – just another day at Microsoft Research Cambridge

Where can you see a Big Data analytics platform that runs in a browser, user interface developments beyond Kinect, and software for programming living cells? At  Microsoft’s Research laboratory in Cambridge – which opened its doors to us on Wednesday this week.

Microsoft chose thirteen different research subjects to demonstrate its approach to the future, at an event called Research Next. Here are the projects we had time to look at all of which are technology that could be commonplace in five years.

Bigger Data

The lab in Cambridge was the first research facility opened by Microsoft outside the US. During its fifteen year history, it has developed technologies that have been applied across the whole spectrum of Microsoft products, from Xbox and Bing to Azure and its data centres.

In its mission statement, the Cambridge lab says it aims to “advance the state of the art in Computer Science”. It isn’t focused on immediate goals and developing solutions for the existing market, opting instead for a “sustainable” approach with a long-term plan and a strong presence in the academic community.

A lot of the current work done by the lab revolves around Big Data. For example, project Rhea – a Hadoop plugin which automatically filters the data sets so only the data necessary for computation would reach the processor, or F#, the programming language based on the .NET framework and designed for analytics.

Even Microsoft’s Xbox 360 gaming platform taps the Big Data potential, to recommend users the games they are likely to enjoy, and answer important questions like “what are the odds that users will like Halo 4?”

According to Drew Purves, senior scientist at Microsoft Research, the biggest markets for Big Data have not yet been discovered. He thinks that in the future, information about the planet as a whole is likely to drive decision- and policymaking at the highest level. With this in mind, Purves and his colleagues are putting together the Madingley Model – an effort to virtually map every single multi-cell organism on the planet, conducted in partnership with the United Nations Environment Programme.

While keeping the decision makers informed is a clear priority, Purves also wants to make Big Data analytics accessible to the masses. To this end, he and his team are working on a yet unnamed project that would allow users to conduct analysis of thousands of variables in a browser window.

Hands-free 

Natural user interfaces is another field that has seen a lot of attention from Microsoft Research. It was in Cambridge that the body part recognition technology used in the hugely successful Kinect controller was born, and the Computer Vision scientists believe that gesture recognition will become the basis for the user interfaces of the future.

At the moment, Microsoft is working on a project called ‘Digits’ – essentially a bracelet fitted with several cameras, sensors and an infrared laser, which is capable of reading minute movements of the hand and convert them into 3D. The technology is a step forward from ‘data gloves’, and during the presentation it drew inevitable comparisons with the interface seen in Sci-Fi blockbuster Minority Report.

Digits was specifically designed to draw very little power and be easily reproducible – the case of the unit on display was made using a 3D printer. Just like with Kinect, we can expect the technology to be applied in video gaming, before more serious applications are considered.

 

Meanwhile, ‘Kinect Fusion’ aspires to use the infrared camera of the Xbox 360 accessory to produce accurate 3D scans of physical objects, and enable free interaction between real and virtual worlds. The system has been tested and it works surprisingly well, considering that it relies on a comparably cheap piece of equipment found in half of the living rooms across the country.

Another project that features Kinect is KinEtre – a software development that can give users complete motion control of any virtual avatar, whether human, a horse or a chair. Traditionally, the process called ‘mesh animation’ is resource-intensive and requires professional expertise, but the Microsoft platform can accomplish the task in real-time. We saw a realistic-looking chair playing football on the screen, mimicking the movements of its operator.

According to research scientist Shahram Izadi, the next challenge for Microsoft is to make the computers aware of their environment, and embrace 3D interfaces, as opposed to the flat grids of icons. He is also full of enthusiasm for augmented reality – a concept that was taken to the next level last year, when the Cambridge lab presented its ‘Holodesk’. Izady says that with recent advances in projection and 3D scanning, a fully-featured Star Trek-style ‘holodeck’ cannot be far behind.

Open Source robots

Among the projects that have left the lab and are doing well in the outside world, special mention goes to .NET Gadgeteer. This open-source toolkit for building electronic devices can be accurately described as ‘Lego for geeks”. It uses the .NET Micro framework and either Visual Studio or the free Visual C # express, and enables creation of a countless number of gadgets, assembled from basic modules like mainboards, cameras, joysticks, buttons, sensors and motors.

Gadgeteer instantly draws comparisons with the open-source microcontroller Arduino. Unlike Arduino, however, working with Microsoft’s platform is completely solder-free. Every module is equipped with a universal socket and is 100 percent compatible with the rest. Even though the technology has been developed by the company, the modules are made by a number of independent manufacturers. The platform has huge educational potential, and has been adopted by several schools and universities.

 

There were other projects worthy of mention, such as the attempts to develop software to program living cells, or Microsoft’s work around Wi-Fi over Narrow Channels (WiFi-NC) that aims to make efficient use of white spaces. We were unable to cover them all, but came away with a strong sense that there is a lot going on at Microsoft in Cambridge.

Look out for our interview with Ken Woodberry, deputy director of Microsoft Research Cambridge, early next week!

Test your Microsoft knowledge! Take our quiz!