# The cornering sequence and driving lines: minimum speed, apex placement, V vs U corners

> Every corner is a sequence: braking, turn-in to apex, apex to exit. Use the corner-radius and curvature channels to locate the apex, judge the line, and decide whether a corner should be driven V-shaped (rotate and drive out) or U-shaped (carry speed).

Area: Driver Analysis · Starting level: L1 Foundation · Updated 2026-11-16
Source page: https://race-engineer-academy.vercel.app/lesson/driver-cornering-sequence-lines

## L1 · Foundation

**The cornering sequence (Segers §7.1):**

1. **Braking point to turn-in** – straight-line braking; where braking starts fixes where the car can turn and how fast it arrives.
2. **Turn-in to apex** – trail braking that blends into a brief, possibly near-zero, neutral-pedal phase (not the same as coasting) to carry speed to the apex.
3. **Apex to exit** – the driver goes to full throttle and onto the next straight.

**Why exit matters most.** Speed out of a corner that leads onto a straight is carried all the way down the straight. Lines should be judged primarily on corners *followed by significant acceleration zones*.

**Minimum corner speed** for a given radius $R$ and grip $\mu$ without downforce:

$$V_{max} = \sqrt{\mu g R}$$

Bigger radius → higher speed, so the fastest line through an isolated corner is the one with the largest radius. But the best line overall depends on what comes before and after.

**Speed-trace shapes.** Look at the speed trace through a corner:
- **V-shaped**: brakes late and deep, short rotation, low minimum speed, early and hard acceleration.
- **U-shaped**: brakes earlier/lighter, carries a high minimum speed through a long constant-speed arc, accelerates later.

Neither is universally right; it depends on the car and the corner.

## L2 · Practitioner

**Finding the line in data.** Without GPS, compute the instantaneous path radius and curvature:

$$R = \frac{V^2}{G_{lat}\, g}, \qquad \kappa = \frac{1}{R} = \frac{G_{lat}\,g}{V^2}$$

The apex is where radius is smallest (curvature peaks). Three patterns (assuming similar track width at entry and exit):

- **Mid-corner apex**: constant-radius arc, symmetric radius trace, lateral g plateau. Typical of constant-radius corners.
- **Late apex**: late turn-in, most rotation early, minimum radius in the first part of the corner, lateral g peaks early then decays; car straightened early for exit → earliest throttle. Costs entry speed. Standard for corners onto long straights and decreasing-radius corners.
- **Early apex**: early turn-in, minimum radius late, lateral g peaks late, throttle late because the car is still turning near the exit. Usually a mistake except in increasing-radius corners or where the next corner demands it.

Hairpins (> 120°) are usually a single late apex, sometimes a double apex depending on how well the car rotates.

**Curvature as a line metric.** Curvature measures how much the *car* turns (not how much the driver steers, which includes balance). For an isolated corner, a lower curvature peak means a wider line and higher potential speed ($V \propto \sqrt{R}$). Average curvature per lap and its scatter compare line efficiency and consistency between drivers.

**V vs U in practice.**
| Favour V-shape | Favour U-shape |
|---|---|
| high power-to-grip (GT3, TCR, heavy cars) | high grip/downforce relative to power (F3, FR, F4 in fast corners) |
| slow corners onto long straights | fast sweepers, corners onto short straights |
| good braking stability, strong traction | cars that lose time when rotated aggressively |

**Use the corner-speed calculator** to see how radius, grip and downforce set the minimum speed, then compare against the logged minimum.

## L3 · Advanced

**Minimum speed with downforce.** With downforce $\tfrac{1}{2}\rho C_L A V^2$:

$$m\frac{V^2}{R} = \mu\left(mg + \tfrac{1}{2}\rho C_L A V^2\right) \;\Rightarrow\; V_{max} = \sqrt{\frac{\mu g R}{1 - \frac{\mu \rho C_L A R}{2m}}}$$

As $\mu\rho C_L A R/(2m)$ approaches 1, the cornering speed becomes limited by other factors (the car is "flat"). For high-downforce cars, larger-radius lines yield proportionally more than $\sqrt{R}$ because the extra speed brings extra downforce.

**Time-based reasoning.** A V-line gives up minimum speed (and some mid-corner time) to straighten the car earlier, so more of the friction circle is available for drive. It pays when the gain on the straight exceeds the loss in the corner:

$$\Delta t_{straight} \approx \int \left(\frac{1}{V_U(s)} - \frac{1}{V_V(s)}\right) ds > \Delta t_{corner}$$

Two questions decide the sign.

1. **What is a km/h of exit speed worth?** If the slower car simply follows the same speed–distance curve shifted back, an exit-speed advantage $\Delta V_e$ is worth about
$$\Delta t \approx \frac{\Delta V_e}{a_x(V_e)}\left(1 - \frac{V_e}{V_{end}}\right)$$
by the end of a straight reaching $V_{end}$. It scales with $1/a_x$: the weaker the acceleration, the more every km/h – of exit speed *and* of minimum speed – is worth. The speed *gap* closes in different ways: when traction-limited (constant $a_x$), $\Delta V \approx \Delta V_0\,V_0/V$; when power-limited ($a_x = P/mV$), $\Delta V \approx \Delta V_0\,(V_0/V)^2$. The gap closes faster when power-limited, but the car is slower there, so the *time* value is not smaller.

2. **Can the V-line create exit speed at all?** Straightening early only converts into acceleration while the car is **traction-limited**. Once it is **power-limited**, extra longitudinal grip goes unused. The crossover is near $V^* \approx P_{wheel}/(m\,a_{x,trac})$.

A GT3 (~270 W/kg at the wheels, ~0.9–1.0 g of traction) stays traction-limited up to ~100–110 km/h, so after a slow corner the V-line turns grip into real exit speed. An F4 (~180 W/kg) is power-limited from ~70 km/h – almost from the apex of a hairpin – so the V-line buys little exit speed while the lost minimum speed is expensive, because its low acceleration makes every km/h costly. Hence U-lines for low-power, high-grip cars.

**Compound corners.** The best line through corner A may compromise corner B; analyse the whole sequence with segment times bounded on the straights before and after. A "slower" entry into a chicane that allows full throttle earlier at the second part is often optimal.

**Consistency of line.** Track the curvature peak and its location per lap: a pro varies apex location by a few metres; large variation indicates searching for the line or reacting to balance. GPS lines (same session) and synced onboard video complement the curvature channel – video is still the best way to discuss kerb usage and reference points with the driver.

**Grip changes.** Rain, rubber, kerb use, banking and bumps shift optimal lines – in the wet, drivers often use off-line grip and smaller curvature changes.

## L4 · Expert

**Minimum time is an optimal-control problem.** The geometric line (largest constant radius, apex at the geometric centre) maximises *minimum speed* for an isolated corner. Lap time is $T = \int ds/V$, minimised subject to the vehicle dynamics, the track edges and the car's g-g-V envelope. Quasi-steady-state lap simulations fix the line and find the speed profile; full optimal-control solvers (Casanova, Sharp & Symonds; Perantoni & Limebeer) optimise line and controls together, including transient yaw.

**What the solutions show.**
- **Max radius ≠ min time.** The optimal line usually turns in later and tighter than the geometric line before a straight: a smaller radius early lets acceleration start sooner, and exit speed is worth more than entry speed.
- **The car rides the envelope.** On a minimum-time lap the combined acceleration sits on the g-g-V boundary almost everywhere – trail braking blends into cornering, cornering into drive. That is the theoretical basis for the combined-g valley diagnostic: any point inside the envelope is time left on the table.
- **The envelope is asymmetric and speed-dependent.** Braking can reach 2–4 g in a downforce car, drive only 0.3–1 g and falling with speed; downforce grows the envelope with $V^2$. The optimal trade-off therefore differs by corner speed even in one car.
- **Transient models vs QSS.** Full optimal-control solvers include yaw dynamics, so they capture the steering and throttle inputs a real transient needs. Casanova, Sharp & Symonds found that changing yaw inertia has a surprisingly small effect on the minimum time itself in a double lane change: the transient model matters more for realistic control inputs and driveability than for the time gained.

**Using it trackside.** Overlay the optimal-control line and speed trace on logged data, and compute envelope utilisation per corner: $G_{comb}/G_{env}(V)$ sample by sample. Where a driver sits consistently at 90 % of the envelope in one phase, that is the coaching or setup target.

**Limits of the theory.** The optimiser drives a perfect, risk-free driver on a deterministic tyre model with no thermal state, no bumps and no consequences. Real drivers keep margin where walls are close or the surface is bumpy, and tyre temperatures change through the lap. Treat the optimal line as a reference that shows *where* time might be, not as a line to copy metre by metre.

## Key formulas

**Maximum cornering speed (no aero)**

$$V_{max} = \sqrt{\mu g R}$$

- \mu: tyre friction coefficient
- R: path radius (m)

**Maximum cornering speed with downforce**

$$V_{max} = \sqrt{\frac{\mu g R}{1 - \frac{\mu \rho C_L A R}{2m}}}$$

- \rho: air density (kg/m³)
- C_L A: downforce coefficient × area (m²)
- m: mass (kg)

**Path radius and curvature from data**

$$R = \frac{V^2}{G_{lat}\,g}, \qquad \kappa = \frac{G_{lat}\,g}{V^2}$$


**Time value of an exit-speed advantage**

$$\Delta t \approx \frac{\Delta V_e}{a_x(V_e)}\left(1 - \frac{V_e}{V_{end}}\right)$$

- \Delta V_e: exit-speed advantage (m/s)
- a_x(V_e): available longitudinal acceleration at exit speed (m/s²)
- V_{end}: speed at the end of the straight (m/s)

**Traction/power crossover speed**

$$V^* \approx \frac{P_{wheel}}{m\,a_{x,trac}}$$

- P_{wheel}: power at the wheels (W)
- a_{x,trac}: traction-limited acceleration (m/s²)

## Trackside takeaways

- Judge lines first on corners followed by long acceleration zones – exit speed is carried down the straight.
- Minimum radius (curvature peak) locates the apex; early peak = late apex, late peak = early apex.
- Late apex for corners onto straights and decreasing-radius corners; mid apex for constant radius.
- V-shaped corners suit high power-to-grip cars in slow corners (they stay traction-limited on exit); U-shaped suit high-grip, lower-power cars and fast corners.
- Analyse compound corners as one sequence with segment times bounded on straights.

## By car category

### GT3

High power and mass, ABS and good traction: V-style in slow corners is usually fastest – brake deep, rotate, straighten early, full throttle.

### TCR

Front-drive with limited traction: getting the car rotated early (often with trail braking and lift-off rotation) to straighten for exit is critical; V-style with a strong exit focus.

### Formula 4

Little downforce, modest power: carry minimum speed (U-shape) in most corners; over-slowing costs time that the engine can’t win back.

### Formula Regional

Downforce makes fast corners about commitment and minimum speed; slow corners still reward a clean exit.

### Super Formula

Very high grip-to-power: minimum speed in medium/fast corners dominates lap time; U-style with high-commitment entries.

### Hypercar (LMH/LMDh)

On LMH 4WD cars, front-axle deployment only starts above a BoP-set speed (typically 150–190 km/h), so it adds AWD traction on fast-corner exits only; slow-corner exits are still RWD and traction-limited. LMDh hybrids drive the rear axle. Total power is BoP-capped, so the hybrid changes how drive is distributed, not how much power there is.

## References

- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §7.1 The Cornering Sequence
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §14.7 The Driving Line (§14.7.1–14.7.4)
- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 1 The Problem Imposed By Racing (minimum time and path)
- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 9 "g-g" Diagram (acceleration-limit boundary, traction vs power limit)
- Other: Casanova, Sharp & Symonds, "Minimum time manoeuvring: the significance of yaw inertia", Vehicle System Dynamics 34 (2000)
- Other: Perantoni & Limebeer, "Optimal control for a Formula One car with variable parameters", Vehicle System Dynamics 52 (2014)
