Theory, Physics and OMG, Surely It Isn’t That Complex?
Despite rubber tyres, a car has a finite amount of grip. Engineers and simulators don’t just map out a “smooth curve.” Instead, they treat the track as a coordinate system and use physics engines to find the absolute limits of traction.
The process relies on several core principles:
1. The Physics Foundations
To calculate the line, software uses specific mathematical equations derived from physics:
- The Radius Factor (r): The maximum speed a car can carry through a turn is proportional to the radius of its arc. A larger radius requires less steering angle, lowering centripetal force and allowing higher speeds.
- The Friction Circle (Grip): A tyre has a maximum threshold of total grip. It can use 100% of it for braking, 100% for cornering (lateral), or a combination of both (e.g., 50% braking and 50% cornering). The line must be plotted to ensure the combined forces never exceed this “circle” of grip.
2. Factoring in the Car’s Characteristics
The hypothetical line shifts drastically depending on the vehicle’s specific parameters:
| Car Characteristic | Mechanical Effect | Plotted Line Result |
| High Power-to-Weight (e.g., F1) | Can accelerate incredibly fast out of corners. | Late Apex (“V-shaped” or Diamond line): The car sacrifices entry speed to pivot sharply and straighten the exit early, getting the power down sooner. |
| Low Power-to-Weight (e.g., Go-Karts, Miatas) | Lacks the horsepower to recover lost speed quickly. | Geometric Apex (Smooth Arc): The line focuses entirely on maintaining rolling momentum, prioritizing the widest possible radius. |
| High Aerodynamic Downforce | Grip increases dramatically at higher speeds. | Higher Minimum Speed Lines: The car can tolerate tighter radii at high speed because the air pushes it down, allowing more aggressive entries. |
Notice they cause very different lines.
3. The Computational Plotting Method
To map the line automatically, modern simulation programs (like those used in F1 or advanced sim racing) use Optimal Control Theory or genetic algorithms (notice something here?). They follow these steps:
- Track Discretisation: The track boundaries are translated into a mesh grid of thousands of sequential points.
- Initial Pathing: A baseline path is established (usually the geometric center line).
- Iterative Solving (Simulated Annealing): The computer runs thousands of virtual laps. With each pass, it nudges the car’s path slightly to the left or right at every point along the track.
- Time Evaluation: If a shift results in a quicker sector time, that modification is saved. The algorithm continues to warp the path until it finds a sequence where changing the line any further results in a slower lap time.
Ultimately, the perfect hypothetical line balances the shortest distance against the highest speed path, squaring off tight corners to maximize upcoming straightaways while smoothing out sweeping curves to conserve momentum.
My first neural network app essentially built a (basic) bespoke genetic algorithm-based vehicle dynamics simulator from scratch.
It shows the emergence of a clean, physics-bounded path through sheer survival of the fittest, which is exactly how top-tier motorsports engineers and simulation software firms validate their mathematical models. Who’d have thought?
Apparently, in the industry, the “known approach” is categorized under evolutionary trajectory optimization. While my neural network used a generational trial-and-error approach to discover the physics, professional engineering firms approach it from two main parallel paths.
