Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - ROCSAFE (Remotely Operated CBRNe Scene Assessment Forensic Examination)

Teaser

The overall goal of ROCSAFE is to fundamentally change how CBRNe (Chemical, Biological, Radiation/Nuclear, explosives) events are assessed, in order to and ensure the safety of crime scene investigators by reducing the need for the investigators to enter high-risk scenes when...

Summary

The overall goal of ROCSAFE is to fundamentally change how CBRNe (Chemical, Biological, Radiation/Nuclear, explosives) events are assessed, in order to and ensure the safety of crime scene investigators by reducing the need for the investigators to enter high-risk scenes when they have to determine the nature of threats and gather forensics. This will help to protect human life, reduce threats, and ensure that the crime scene is processed in a more efficient manner.

For this, ROCSAFE will make use of cost-effective modern remotely-controlled robotic air and ground vehicles (RAVs/RGVs) that are designed for use in rain, wind, and challenging ground surfaces and obstacles. First, RAVs will assess the scene. These will have cameras and can carry an array of innovative new high-performance and rugged miniaturised sensor systems for RN, chemical and biological threats.

To reduce the crime scene manager’s cognitive load, ROCSAFE will include new Central Decision Management software and a Command Centre. All images and data will be streamed to this, where it will be analysed and displayed on a sophisticated and intuitive interface with maps and video, showing results of analytics and giving readings geographical context. This will enable the scene commander to assess the nature of threats, develop an Action Plan and an Evidence Plan, supported as needed by the Central Decision Management. It will also assist in coordinating sensors and mobile units. Its data analytics will provide fusion of multiple sensor data sources, to allow probabilistic reasoning about the most likely threats and likely locations of epicentres.

After the scene is assessed, RGVs will be dispatched to collect forensic material/evidence, with automatically-optimised routes to avoid hazards. They will have innovative new equipment for forensics collection that will automate best practices. Forensic material will be collected, bagged, tagged, documented, and stored by the RGV.

Overall, ROCSAFE will ensure that CBRNe scenes are assessed more rapidly and thoroughly than is currently possible, and that forensic evidence is collected in a manner that stands up in court, without putting personnel at risk.

Work performed

The first 18 months of the project have focused on requirements, design and prototyping.

Detailed operational scenarios have bene formulated, along with end-user requirements.

Partners are developing new CBRN sensing systems that can operate with aerial vehicles and ground-based robotics. During the first reporting period of ROCSAFE, work has been done on design, building and preliminary testing of chemical sensors: the lightweight self-limiting pre-concentration device for the aerial vehicle and the pre-concentration and FAST-GC separation module for the ground vehicle. Both modules have proven to work successfully with the ROCSAFE target chemicals in laboratory conditions. For biological detection, design and development of a lab-on-a-chip device is underway. Lightweight robust RN detectors have been specified and their development and characterisation is underway.

Work is also underway on the central decision management system and graphical user interface. A communications protocol has been developed, based on past work in other research projects, to allow all components to exchange information securely. A virtual world simulation has been built as a test-bed, and this is being used to evaluate artificial intelligence methods for object detection and multi-agent system algorithms for coordination of a swarm of aerial vehicles. Initial work has started on combining information from multiple sources using probabilistic reasoning, and ranking documentation by relevance. In addition, work has started on the design of the graphical user interface.

Aerial and ground vehicles are being adapted for ROCSAFE, based on existing commercial platforms. For the aerial robots, a sensing/sampling cluster termed a “turret” has been designed, that can be lowered to operate without being affected by the down-draft from the vehicle’s propellers. A self-levelling landing pod is being designed that the aerial vehicle can land on to deposit samples. This is towed into place by the ground vehicle, to position it on the edge of the warm zone. In addition, new tools are being designed for the ground vehicle to facilitate collection of multiple kinds of forensic samples. An RFID tagging mechanism has been developed for the forensic samples.

In addition to all of the technical work described above, the participants have developed communications materials, a data management plan, and an exploitation plan. The consortium has participated in a total of at least 15 dissemination events in the first 18 months.

Final results

The lightweight chemical sensing system under development for the RAV is a novel combination of a MEMS-based sampling and pre-concentration device coupled to a wide-range IR spectrometer, resulting in a small and quick sensing device with chemical identification capabilities. It is expected to have applications in several fields beyond CBRN, in particular because of the relative simplicity of the device allowing for a low-cost sensing system. Industrial safety, air quality and contamination, as well as food quality can be addressed.
The C sensor for RGVs under development is the fist example of a QEPAS analyser hyphenated to a FAST-GC module, thanks to a miniature interrogation cell with a few microliters internal volume, to allow fast and sensitive analysis of trace mixes on the field. The FAST-GC module is a more complex and expensive equipment, and miniaturization is foreseen as a follow-up activity beyond ROCSAFE.
Swarm-based routing algorithms for the RAVs and RGVs will optimise RAV and RGV navigation, to maximise the utility of sensor readings and appraise the scene as thoroughly as possible and collect relevant forensic evidence efficiently, while minimising disruption of the scene.
Probabilistic reasoning and intuitive graphical displays will reduce cognitive load for the crime scene personnel, by providing intelligent decision support to assess risks, retrieve relevant documentation when needed, and displaying all information in a clear manner.
Deep learning will support decision-making by automatically identifying relevant objects and drawing attention to parts of images and videos that contain anomalies or other important features.

A UAV will be developed according to the requirements necessary to evaluate the scene. Its semi-autommous flight capability, including the avoidance of any obstacle, will allow to mitigate the personal damage caused by a CBRNe attack and will facilitate and speed up the work carried out by the personnel.
All these innvations should make it possible to improve, detect and assess any CBRNe attack with the consequent social impact that this entails.

Website & more info

More info: http://www.rocsafe.eu/.