You are Home   »   News   »   View Article

LR - defining "SafetyTech"

Tuesday, September 22, 2020

Lloyd's Register has a mission to establish 'SafetyTech' as a technology category similar to 'FinTech' - hopefully encouraging large amounts of investment and innovation into the sector. It showcased some interesting companies in an event in London in October.

Risk and technical services organisation Lloyd's Register has a mission to set up 'SafetyTech' as a class of technology, similar to 'FinTech', hopefully driving innovation and investment in technology to improve safety.

It held a 'Safety Accelerator Showcase' event in London in October 2019, as an opportunity for delegates to get immersed 'in digital safety and risk innovations and entrepreneurship.'

LR's 'Safety Accelerator Program,' aims to match corporates which have specific safety technology needs with technology companies. Corporates can set challenges, or specific problems they would like solved. Technology companies are invited to submit proposals for how they will solve them. The Lloyd's Register Foundation gives a grant to pay for the pilot project.

The Safety Accelerator is run in partnership with Plug and Play which claims to be 'the largest global innovation platform and most active venture capital fund in Silicon Valley'. It is then able to invest in the companies.

'We end up with an ecosystem and a community. tech company and corporates,' said Steve Price, LR Accelerator Programme Lead. 'Our job is to introduce people.'

One idea for safety technology could be an analytics tool which can help company safety departments get more insights from incident data reports. There can be too many for a safety manager to read them all, so perhaps they could just see the trends.

Another idea for safety technology could be maintenance software which advises on what maintenance does not need to be done - rather than just tell us what maintenance is due.

There are a number of barriers to running pilot projects with safety technology, Mr Price said. It is very difficult to calculate the return on investment in safety technology. Also running pilots can be quite expensive, particularly when the business case is not clear. Projects can easily get de-prioritised behind something which feels more immediately important.

Sensing Feeling

Sensing Feeling (www.sensingfeeling.io) of London makes tools to 'sense' human emotion, using video analytics. It is working together with shipping company Scorpio Group. The idea is that it might be possible to analyse how seafarers are feeling.

The company did not initially envisage that shipping would be a market for the technology, said Jag Minhas, CEO and founder. It was anticipating more interest from sectors like retail, events and hospitality, where companies might have an interest in sensing how people feel.

For Scorpio Tankers, the company is aiming to create a 'risk index', mapping seafarer emotional state to risk of giving them control over a vessel, collecting data from real and simulated voyages.

There are questions of how relevant a model developed for one person would be to another person, since every individual is different. The system does not collect any personally identifiable data, for confidentiality reasons. But it may be able to adapt the models to take into account traits of people with certain gender or culture, the spokesperson said.

Ohalo

Ohalo (www.ohalo.co), based in London and Sunnyvale, California, is developing a tool to automatically anonymise health and safety data, so it could be useful for further analytics studies without violating anyone's confidentiality.

The UK Health and Safety Executive has 1.5m accident reports, and would like to share this data with data scientists, to see if they can spot trends in it, said Kyle DuPont, CEO and co-founder.

To anonymise the data manually was estimated to take 70 man years.

The system can automatically identify data such as name, company and address, using custom data classification models. It can then automatically remove (redact) it.

There are options about how much redaction is most appropriate, for example it may be useful to the data analytics to have some geographical context, and a decision can be made that this is not too much of an issue with confidentiality.

The auto redaction has been compared to manual redaction and is proven to be more accurate in some places, for example the computer may spot first names in the text more accurately than a person does.


M Squared Lasers

M Squared Lasers of Glasgow, UK, is developing laser technology which can analyse food condition when it is in a sealed bottle. Light interacts with different substances differently, getting reflected and absorbed in different ways, so a model can be built of how food in different conditions will interact with the laser in different ways. The laser can be 'tens of metres' away from the bottle.

It helps that the company 'makes one of the world's purest lights', said Robin Head, innovation laser engineer.

Senseye

Senseye of Austin, Texas (https://senseye.co/) is analysing the way that iris muscles (which control the size of our pupils) are connected to our minds. Scientists have long known that our pupils become larger (dilate) for many different reasons, which is interesting but not so useful for analysing the brain state. But if you look at the muscles which pull the iris (which controls the size and diameter of the pupil), rather than the pupil itself, you can get much more insight, says David Zakariaie, founder of Senseye.

The company aims to measure individual muscle fibres.

A trial is being run with Pacific International Lines, a container shipping company based in Singapore, where crew are asked to take an 'iris test' before starting work. In theory, if the system detects a reduced psychological state (for example from fatigue), a seafarer could be assigned to a less dangerous job.

The system was installed in a trial running from June to October 2019.

It developed a training model, giving people slowly increasing amounts of alcohol to drink in a laboratory, verifying blood in their system using a breathalyser, and seeing how that affected their iris muscles. But the ambition is that the system will be much more than an expensive breathalyser, it will be able to assess people's psychological state.

The head of safety at Pacific International Lines is a former investigator at Singapore's Transport Safety Investigation Bureau, Mr Zakariaie said. In this role, he observed that a number of accidents occurred because of the individual was not in a state of mental fitness. This mental fitness state was then linked to a specific cause, such as bad news from home.

There are three stress factors where most people's irises respond in the same way - alcohol, drugs and fatigue. But for more complex stress factors, individuals respond differently - so you would really need to develop models about individual people to identify their mental state, he said.

In this, the computer's capability is similar to a human's - we could probably detect a stranger's alcohol intake, drugs intake, or fatigue by looking into their eyes if it was extreme enough, but we would need to know them personally to recognise their mood from looking into their eyes.

Another challenge is that the computer modelling uses a large amount of computer power - which can be hard to make available onboard a ship, Mr Zakariaie said.

Senseye has also worked with the military, testing the iris muscles of troops after and after deployment.

Allergy Amulet

Allergy Amulet, based in Wisconsin, is developing a testing system which people with allergies can use to test out food where they don't have complete information about the source, such as in restaurants.

It has developed a small electronic device, which analyses a sample of food.

Ms Barnes's motivation for developing the company comes from having a food allergy herself and having a 'near death experience' accidentally consuming food containing an allergen. She is a lawyer and her co-founder is a chemistry professor.

In the US, 1 in 13 children and 1 in 10 adults have a food allergy, she said. And 90 per cent of allergies come from just 8 foods.

For someone eating in a restaurant, the only allergy information normally comes from asking the serving staff, and 'they don't always get It right'.

LexaTexer

LexaTexer of Berlin is looking at ways to do better analytics on incident reports. The company specialises in unstructured data.

'It isn't possible for a human to read all the reports and they may miss something,' said Guenther Hoffman, company founder.

It has already done business with a number of car part suppliers, helping them learn from past incidents.

One idea is to look for data about the environment certain accidents occurred in, such as rain or an earthquake. By spotting trends, safety managers may identify that work in certain environments is too hazardous.

The analytics could also be used to spot patterns, such as certain body parts proving at much higher risk of injury than others, or certain accidents happening at certain times of year, or certain trends in the incidents.
Ultimately it may be possible to correlate someone's level of training with a reduced likelihood of an accident. This would help justify the investment in training, he said.

Invision AI

Invision AI of Toronto (www.invision.ai) is looking at better ways to process data within sensors, so that only important data needs to be communicated, rather than all of it. The jargon is 'processing data at the edge', says CEO Karim Ali.

The company is developing a video analytics system which can determine how many participants are in a car, with a live video image of a car travelling at 'highway speed'. It can be used to enforce regulations where a car must have 2 or more occupants to use a certain lane.

If the entire video stream was sent to the cloud for running analytics, it would take too long - so instead, the analytics is done within the roadside camera.

It has a project with BS Shipmanagement to use video analytics to identify if the oily water separator is being properly operated. This can include facial recognition, to check that the person operating the equipment has authorization.

There is no need to upload any personal data to head office, just data that for example someone is operating equipment without authorization.



Associated Companies
» Lloyd’s Register
comments powered by Disqus

CREATE A FREE MEMBERSHIP

To attend our free events, receive our newsletter, and receive the free colour Digital Energy Journal.

DIGITAL ENERGY JOURNAL

Latest Edition May-June 2021
May 2021

Download latest and back issues

COMPANIES SUPPORTING ONE OR MORE DIGITAL ENERGY JOURNAL EVENTS INCLUDE

Learn more about supporting Digital Energy Journal