The term personal protective equipment (PPE) refers to devices with the function of safeguarding the person who wears them, or in any case carries them, from risks to health and safety. There are different kinds of PPE, linked to the worker's activity, and protect them from multiple risks intrinsic to their work. The most frequently used PPE in healthcare is for respiratory, conjunctival, hand, body, feet and head protection. And let's not forget that, with COVID-19, we have changed our habits by wearing protective masks as PPE every day. Here are the best technologies for checking that workers are making use of PPE.
How to check the effective use of PPE
Once the need to increase personal protection has been ascertained, it's then necessary to check/verify that PPE is actually being worn. If we take the example of masks for all access to premises, the safety officer must check that people are actually wearing them. Especially when it comes to industrial environments, the correct use of PPE is essential for any company to reduce the number of accidents. If we talk about numbers, the United States Department of Labor estimates that the annual costs associated with occupational accidents and diseases amount to $170 billion, a very high figure that clearly shows how fundamental prevention and control are. Technology comes in handy for checking PPE.
Bluetooth and UHF RFID technologies
The technologies used to check PPE include Bluetooth, which simply involves applying a BLE tag to each item of PPE and using a wearable transponder that detects the signals from each item, in order to transmit it back to the central system to process any alarms. This completely effective solution has significant limitations that only make it suitable for a limited number of people, because:
- the PPE must be certified with the associated tag (limited choice of purchase of PPE and its type)
- battery management: in many cases it is preferable not to change the battery, but to buy new PPE with an associated tag; we also have to take into account that PPE is subject to expiration and wear
- recharging the transponder: depending on the technology used, daily recharges are also required
- it is not possible to apply a tag to all PPE (e.g. hearing protectors)
- the PPE costs, which increase considerably since they must have the BLE tag integrated
Based on these considerations, Bluetooth technology is suitable for companies that need continuous checks on a small number of workers, as the commissioning costs are very high. The Bluetooth solution is not scalable.
There are other solutions on the market to monitor PPE, like the adoption of UHF RFID technology. This is scalable compared to Bluetooth technology, but requires expensive proximity readers and gate arrangements for checks.
Cameras and artificial intelligence
The availability of local systems capable of processing images quickly and the use of increasingly fast and reliable algorithms make artificial intelligence the best technology for recognition and correct use of PPE. The previous systems, as well as being difficult to implement, are limited to verifying the presence of PPE but struggling to verify its use (e.g. tags can confirm whether a helmet is present, but not whether it is worn correctly; the worker could be holding it and the system would not be able to tell).
The use of cameras makes it possible to survey video streams in all areas of interest, which will be processed locally and can inform the safety officer of non-compliance in real time, also notifying the worker. It is essential that the image processing is performed locally and no image is stored but only the outcome of the processing, in order to fully comply with issues of privacy management. Another important aspect is that everything must be managed anonymously in order to comply with the regulations on monitoring workers.
Different technologies are required to meet this need: as regards cameras, the classic IP cam is recommended, while AI processing requires the use of gateways equipped with GPU. Cameras are also emerging on the market with a GPU card already integrated, which further simplifies the installation process. As for the artificial intelligence algorithms to use, there are several on the market: the following diagram shows the speed of execution. Among the best-performing algorithms, YOLOv4 certainly stands out.
The use of artificial intelligence has the considerable advantage of enabling modifications without modifying the infrastructure. If we think about the adoption of new PPE, for example, we only need to train the network to recognize it. Conversely, with tag technology we have to buy, in the best case, or create, in the worst case, new PPE with an integrated tag. The network can be trained not only to recognize PPE but also to verify its correct use, and can evolve over time according to the needs of workers and employers.
One last aspect to take into account: the infrastructure thus created, as well as checking PPE, can also evolve to monitor certain safety situations such as the presence of workers in prohibited areas or to identify other specific elements of production activity.
PPE has been subject to verification in working environments, in particular industrial ones where the need is pressing, but artificial intelligence can also help with simple checks on the presence of masks in public places, to comply with anti-COVID-19 directives. In this case, it is possible to adopt totems where people must stop to certify that they are wearing PPE correctly.
In conclusion, we can say that artificial intelligence technology is:
- suitable for different environments and markets
and today is probably the best technology for checking and verifying PPE integrated with an Internet of Things architecture for transmitting information.