Navigation and Content

Logistics Use Cases: People - Health and Safety

Health and Safety Applications

Workplace Safety

Warehouses, yards, depots, and other busy logistics facilities represent many risks for the health and safety of a workforce. This is where artificial intelligence comes into play. Computer vision and AI algorithms can help monitor and analyze the movements of people and vehicles including autonomous robotic devices, forklifts, and trucks. A computer vision system can detect speeding, movement in the wrong direction, parking in the wrong place, and more. AI can also be used to enhance workplace safety by identifying real-time noncompliance, such as workers not using walkways, and can issue safety alerts so that immediate action can be taken to minimize risk and reduce unsafe behavior.

With its AI-based unsafe event capture solution, computer vision startup Protex AI collaborated with DHL in a proof-of-concept project to empower EHS teams to take proactive safety decisions. The team succeeded in converting insights into actionable cues to change workflows, improve operational safety, and implement corrective actions. 

Computer Vision in Health - Improving Ergonomic Pose

Human pose estimation (HPE) is a computer vision-based technology that classifies joints in the human body, capturing a set of coordinates for each (pairs of key points) that form a skeleton-like representation of the human body and can describe a person’s posture and motion. This information is essential for ergonomics, the study of people’s efficiency in the working environment.

For example, when a person bends down, they are at the highest risk of back injury. Twisting, turning, lifting things incorrectly, and carrying loads that are too large or heavy are also major hazards. 

Working to reduce workplace injuries and to improve workplace health and safety, tech company TuMeke has developed an AI ergonomic risk assessment platform enabled by computer vision. The user records a smartphone video of an associate doing a warehouse task, such as lifting a box, and this data is then analyzed to provide a summary of risk, highlighting unsafe postures.

Protective Personal Equipment (PPE) For Safety At Work

Employers are typically responsible for providing and utilizing personal protective equipment (PPE) at work such as safety helmets, eye protection, and specialist clothing. It’s important that employees always wear the right PPE correctly to protect their health in the workplace. AI can be used to improve workplace safety since computer vision technology solutions are able to track noncompliance. Some systems can even identify the reason for noncompliance, such as the high-visibility vest being too uncomfortable to wear.

AI systems utilized in the workplace can be trained to identify various PPE types in real time. Analyzing video streams from strategically placed cameras, a system can identify whether associates are wearing appropriate PPE and using it correctly. This helps ensure compliance with work safety protocols and avoid incidents. In addition, the system can detect defective and damaged PPE (a major safety hazard) – something that also helps to keep associates safe and secure on the job.

Using AI In The Workplace For Driver Support

Another use case of computer vision improving health and safety is that it can be used to detect signs of human fatigue during long-haul truck driving. For example, facial recognition technology can identify drooping eyelids and changes in facial expression. This is achieved using a combination of cameras and machine-learning algorithms which analyze images to find patterns that indicate that the driver is tired. The system can alert the driver to take a break and even trigger an alarm to warn other road users about this potential danger.

The same AI technology ensuring workplace safety can be used to detect and issue warnings about correct seat belt use and unauthorized personnel entering the vehicle. In fact, many modern computers vision systems perform multiple tasks simultaneously, such as object detection, facial recognition, and image classification.

A great example of this multitasking is in autonomous vehicles – the vision system can simultaneously detect and classify objects in the environment (such as pedestrians, other vehicles, and obstacles), track their movements, and make decisions based on this information to safely navigate an entire journey. 

Israel-based Cipia promises an added layer of intelligence to the in-car and in-cabin automotive environment. Using edge-based computer vision and AI, the company offers a range of sensing solutions including driver monitoring systems and occupancy monitoring systems for better, safer mobility experiences.

Challenges of Implementation

Challenge 1: By definition, a prediction is never 100% correct. If a machine-learning algorithm were to make a mistake about a person's health and safety situation at work, the cost would be unacceptably high.

Challenge 2: It is difficult for most non-statisticians to fully understand and make practical use of AI confidence percentages and accuracy levels. EHS teams may need help to apply data. 

Challenge 3: Computer vision currently struggles to detect some materials such as glass and other transparent materials, as well as things with a shiny surface and materials that change shape or form.

In the warehouse environment, as many as 20,000 US workers are injured in forklift accidents alone each year and 25% of those accidents happen when a forklift overturns.

FAQs About AI and Workplace Safety

How can AI improve workplace safety in busy logistics facilities?

AI, through computer vision, can monitor and analyze the movements of people and vehicles, including robotic devices, forklifts, and trucks, to detect issues like speeding, moving in the wrong direction, and parking violations, leading to improved workplace safety.

What is the role of computer vision in ergonomic pose assessment?

Computer vision-based human pose estimation (HPE) technology can capture and analyze a person’s posture and motion, helping to identify risky behaviors in the workplace, such as bending incorrectly or lifting heavy loads.

How can AI assist in the workplace by ensuring employees wear the right personal protective equipment (PPE)?

AI-powered computer vision systems can track PPE compliance by analyzing video streams from cameras to identify whether employees are wearing the correct PPE and using it properly, contributing to workplace safety.

How can AI detect driver fatigue?

AI, with facial recognition technology, can identify signs of driver fatigue, such as drooping eyelids, prompting drivers to take breaks and alerting other road users to potential dangers.

What challenges are associated with implementing AI for better workplace safety?

Predictions made by machine-learning algorithms are not always 100% accurate, which can have high costs in terms of health and safety. Additionally, understanding AI confidence percentages and accuracy levels may be challenging for non-statisticians. Additionally, computer vision often has difficulty detecting materials like glass, transparent materials, shiny surfaces, and materials that change shape or form, which could prove problematic when using AI in the workplace.