AI/ML models for high-accuracy, faster multilayer flexible packaging identification

Overview

The proposed solution leverages advanced AI/ML models to enhance the identification and routing of multilayer flexible packaging, using existing infrastructure. This approach targets decision-making bottlenecks in sorting processes, rather than relying on new equipment. The models are intended to reflect actual material behaviors and operational constraints to support scalable, practical use in the packaging industry.

Technical specifications

The AI/ML models are designed to interpret available signals more effectively, addressing current limitations in flexible packaging recovery. This is achieved through:

• Enhanced decision-making algorithms that optimize sorting accuracy and efficiency.
• Pilot studies to measure gains in accuracy, purity, and recovery in realistic settings.
• Alignment with industry evaluation criteria to ensure performance targets reflect real-world conditions.

These features enable significant performance improvements without the need for additional infrastructure.

Technology readiness level

The technology is at TRL 4, having been validated in a laboratory setting and is now being tested in a relevant operational environment through pilot studies. This phase will further assess the model's feasibility, robustness, and scalability, providing a pathway to commercial deployment.