Leveraging Generative Design and Point Cloud Data to Improve Conformance to Passing Lane Layout

Abstract

The inadequate design of highway elements is a major contributing factor to traffic collisions. Guidelines and regulations for highway design serve as a structured framework for engineers, providing essential direction to create compliant models. These rules are established to ensure the safety and well-being of road users, addressing aspects such as road geometry, signage, lane markings, and more. Developing the best and most efficient solution within specific design requirements involves evaluating various choices. Generative design is a process that employs algorithms and computational methods to generate and assess design solutions, aiming to find the most optimal outcome meeting criteria like safety, capacity, efficiency, and sustainability. Building Information Modeling (BIM) involves creating a digital model of a structure, incorporating physical and operational attributes. This thesis employs a user-friendly, logic-based language to describe rules for designing highway passing lanes. The integration of BIMKit enhances the highway design process, streamlining and improving efficiency. Leveraging the faster evaluation process of generative design facilitates the creation of highway passing lanes. The objective is to analyze 16 real-world existing passing lanes in Alberta, demonstrating the method's applicability in the transportation field. Focusing on passing lanes within the transportation system offers a manageable starting point to grasp the challenges of adopting the proposed approach without overwhelming complexity. It allows for a gradual build-up of generative design expertise, enhancing our understanding before tackling broader aspects. Improving conformance to passing lane designs directly impacts road safety and traffic flow, demonstrating the practical benefits of generative design. Traffic sign placement, governed by clear safety and regulatory guidelines, offers a well-defined basis for generative algorithms. The availability of relevant data facilitates accurate model training. This approach serves as a proof of concept for generative design in highway transportation, initially concentrating on traffic signs as a controlled testing ground before expanding to more complex design rules including other passing lane characteristics. The analysis aims to determine the compatibility of the current passing lanes with the design guidelines recommended by the MUTCD and the Highway Geometric Design Guide of the Alberta government. Subsequently, generative designs conforming to code requirements are proposed, ensuring they meet necessary standards and specifications. The examination reveals significant improvements in overall score and guideline adherence through the implementation of generative design and required adjustments. The compatibility of the considered passing lanes increases from an initial average of 61.82% to an impressive 91.31% after leveraging generative design techniques. The results highlight the importance of incorporating generative design in the transportation domain. By utilizing advanced algorithms and computational methods, generative design greatly enhances infrastructure planning and design efficiency, providing a powerful tool for engineers to create innovative and compliant solutions for highway projects.