by Lovelesh Parasher, CTO, Orbit Techsol Group
Safety is one of the prime concerns of all manufacturing units. Companies are under constant pressure to stay competitive, keep costs down, deal with supply chain disruptions, increase productivity, adapt to new technologies, hire and retain skilled workers, and manage countless other realities of doing business. While handling business contingencies, safety can take a back seat and tend tocome to the forefront in case of an adverse incident.
This is where VEDA can be brought in to democratize safety. VEDA is not just a computer vision tool to identify safety incidents. It acts as the interface to fine-tune and direct specific strategies for infringement reductions and the way ahead.
To understand the interplay of safety and VEDA, we will discuss Taloja, a manufacturing unit of Hindalco Industries. Taloja has contract workers working in the plant. These workers are rotated regularly, and although safety training is provided to new entrants as part of onboarding, the plant wants them to adhere to training at all times. Prevalence of unsafe practices was previously not quantifiable as workers would adhere to safety standards while in the presence of a supervisor. However, when left alone, they sometimes would revert to unsafe practices (such as removing their PPE’s). The Business also wanted to increase the efficacy of the Group safety talks as they wanted to understand which unsafe practices or areas to target.
The solution to the aforementioned business case is not just AI detection but also bringing about a positive behavioural change in employees while interacting with all stakeholders in an unobtrusive manner. VEDA broke down this problem into multiple levers and provided a solution to its different aspects.
Here’s an overview ofthe breakdown of the problem:
- Identify safety infringements – Solved in VEDA using deep-learning algorithms to identify a variety of unsafe practices (people not wearing a helmet, masks, reflective vests, etc).
- Constant vigilance – VEDA operates 24x7 without any intervention; no incident goes unnoticed.
- Lower cost of implementation –VEDA can be scaled to monitor multiple cameras and use cases using minimal hardware to minimize costs by optimizing algorithms used by computer vision experts.
- Identifying incidents hotspots – With VEDA, the plant was able to identify areas where specific violations were higher in number and could develop a pointed strategy to solve the issue.
- Disbursing safety responsibility – In VEDA, a person can be assigned responsibility for an area/ type of violation. This makes it easier for the upper management to drive accountability for enhancing the safety of specific areas.
- Rewarding people for safety interventions – As people can be assigned to a specific area, the plant rewarded people who showed significant improvement in safety adherence in their specified area.
- Real-time alerts – In many cases, a slight nudge while a person is walking in a restricted area or removes his helmet goes a long way in inculcating a habit of following safe practices.
Output post implementation:
- In areas where real-time alerts were configured, safety violations reduced by a factor of ~20% over 2-3 months.
- Hotspots for violations were identified during a specific time (lunch). Staggered lunch timings for personnel based on the insight led to a significant reduction in incidents on the platform.
- The discussion in safety meetings enhanced from singular events to specific action to be taken based on trends of violationsacross areas.
- Upper management started to track safety from incident metrics and was able to reward personnel who improved safety compliance in their area.
Manufacturing facilities require the capabilities to drive actions and change management to enhance compliance to safety in addition to the AI-based safety monitoring system. They needed a system that would be easy to interact with, can aid in creating directed safety interventions, have a real-time alerting system and means for upper management to reward people improving safety. In short, the units need an empathic system with AI at its core. This is where VEDA operates and can add tremendously to any organization for driving compliance to safety practices.