A physical crime scene may be the main topic under investigation in a criminal trial, but despite its importance, jurors are often reliant on less-than-satisfactory second- and third-hand information about exactly what took place there.
“The problem with today’s crime scene reconstruction practices is that [they] usually involve still photography, hand-drawn sketches and — in rare cases — videography,” Mehzeb Chowdhury, a PhD researcher in forensic science and criminal investigations at Durham University in the U.K., told Digital Trends. “Experts will bring 3D-rendered crime scene animations, which are later created and rendered using a combination of the still images and sketches, to court. This is an approximation of reality, not reality itself. Juries are bamboozled by conflicting crime scene recreations, as each side presents its own version of the crime scene, and where the evidence was found.”
Chowdhury’s solution? A robot that will enable jurors to explore crime scenes for themselves using virtual reality. His MABMAT robotic imaging system is capable of recording 360-degree HD video using a NASA-inspired rover unit, able to autonomously roam a crime scene at the time it’s being investigated and capture every salient detail from it.
The rover is built on a combination of two low-cost microcontroller boards, Arduino and Raspberry Pi, and runs on open-source software. The VR footage it records can then be viewed at a later date, using any headset down to a low-end smartphone with Google Cardboard headset.
This isn’t the first time virtual reality has been considered as a means by which to let juries take a look at crime scenes. Previous attempts have explored using everything from lasers and video game engines to help model individual crime scenes after the fact, to other efforts involving Hollywood-style green screens. There are two main differences with Chowdhury’s concept, however.
The first is the price point. Even taking into consideration the camera and robot, the entire system costs less than $400. There is also little cost involved with the creation of the virtual reality “scenes” jurors would get to explore.
“If we had gone for the traditional means of VR-content creation, such as 3D scanning a crime scene, texturing it, and then using a gaming engine to render the world, it would lead to impediment in accessibility for our intended users,” Chowdhury said. “Today’s VR-ready computers require massive amounts of processing power, expensive graphics cards, and headsets with eye-movement and head-tracking capabilities.”
The second, perhaps more crucial, point is that Chowdhury’s entire mission statement was — as noted — to take away non-objective guesswork from the courtroom. Had he relied on recreating scenes based on eyewitness reports or CCTV, there would be the possibility of bias being introduced.
“Unlike 3D recreations, [my system] would be true representations of how things were, rather than a user-created propaganda video to sway the jury,” he continued. “The most-problematic aspect of crime scene visits is that, with time, every characteristic of the scene changes in some way or the other. This is called scene degradation. Years could pass between a crime being committed, and a jury scene visit, with very little remaining the same. A contemporaneous snapshot of the entire crime scene would preserve the necessary details for investigation and trial.”
As to how long it is before tools like this can be deployed, Chowdhury suggested that it might be sooner than you think. “Realistically the system is a few months away from field-testing,” he said. “The plan is to work with police departments in the U.K. and U.S. Around 50 police forces from these two countries have already participated with data. The ideal scenario would be to collaborate with them. Unlike other projects which are looking at similar technology, this is self-funded, with months spent in my own garage. The system is on-track for testing in the next few months, but further development would depend on the support it gets.”
Perhaps the biggest question of all is what possibilities this poses for humans in the legal system? After all, we now have Artificial Intelligence being used in policing, AI lawyers being consulted by clients, and the possibility of AI-driven judges in the future. Now all we need is for robot jurors and someone to change the laws so an algorithm can be accused of a crime …
Hey, stranger things have happened!