ENGS 89/90 Reports

Year of Graduation

2021

Project Advisor

Scott Davis

Instructor

Solomon Diamond and Rafe Steinhauer

Document Type

Report

Publication Date

2021

Abstract

Significance: On the Formula One circuits, cars operate at the upper limits of vehicle performance, pushing racing technology to exciting new boundaries. To achieve this, it is critical to understand in detail the inter- actions between the car tires and the track surface itself. As such, Pirelli—the tire manufacturer responsible for selecting tire specifications for each race—has systems in place to measure the macro and micro roughness of the asphalt at the track surface. Formula One itself, however, wants the ability to collect their own mea- surements so that they can provide insight into Pirelli’s decisions and create internal specifications for track roughness required on future circuits.

Objectives: The primary goal of this project is to test the feasibility of using photogrammetric methods to accurately and affordably gather track roughness data to be used by Formula One. This proof­of­concept will lead to a design that improves on the state­of­the­art surface profiling devices currently on the market. To understand the track roughness well enough to predict tire interactions, it is necessary to examine about 20 areas around the track of at least 100 cm2, a size corresponding roughly to the contact patch of a tire. For each patch, the roughness must be examined at a depth range of at least 3 mm below the surface to characterize the portion of the track with which the tire directly interacts. Additionally, the model must be generated at a resolution of at least 0.5 mm, and ideally as close to 0.01 mm as possible. There are some state ­of ­the ­art solutions that exist for this purpose, but they are all either cost-prohibitive, time­ intensive, difficult to transport, or not user­-friendly. Thus, this presents an opportunity for innovation: to improve upon what is on the market and deliver a cheaper, more accessible product to Formula One.

Innovation: Using 3D photogrammetry (images taken from various positions and angles) to solve this problem affords the highest opportunity for success and innovation, while also providing useful data to Formula One. Our final device takes this approach. It includes a Raspberry Pi integrated with three cameras and a motor, which are mounted onto a rotating platform with a circular viewing window cut out of the middle and a ring- light affixed below it, all powered by a portable battery pack. At the press of a button, the device is powered, lighting up the area of interest, and the stepper motor runs, rotating the cameras by small increments and taking a series of photos of the pavement at different angles. These photos are saved to a portable USB drive, which can then be uploaded to Agisoft Metashape, a professional photogrammetry software that analyzes the photoset and generates a 3D model. This device design optimizes efficiency (three cameras take photos nearly simultaneously, enabling multiple different angles to be captured at each rotational step), repeatability (the device functions identically on each run ­through), and affordability (this mechanical setup is much cheaper than laser­ scanners or other photogrammetric setups).

Approach: After electing to pursue photogrammetry as the surface profiling method, there were three main goals—designing the hardware, designing the software, and creating a robust testing and validation plan. Designing our own software was deemed infeasible within the scope of this project, so we have decided to use Agisoft Metashape. For hardware, we first identified a camera that balanced affordability, programmability, and resolution: Raspberry Pi HQ. The physical device was then constructed and the motor was integrated, followed by the multi­-camera setup, with validation tests conducted at each step using an independently designed and machined “roughness plate.” For validation, the 3D models produced by Agisoft serve as an initial visual gauge of accuracy; additional MATLAB scripts were used to numerically assess the error of the generated model—using metrics such as RMSE—against the benchmark plate.

Impact: Our solution demonstrates the viability of using photogrammetry to successfully characterize track roughness. Formula One can build off of and upgrade this device to assess the roughness of their existing tracks and use that knowledge to advise on what is required for future race track surfaces.

Level of Access

Restricted: Campus/Dartmouth Community Only Access

Download

Restricted

Available to Dartmouth community via local IP address.

Share

COinS