String Stability of ACC Vehicles in Mixed Indian Traffic

The problem

Most string stability theory assumes homogeneous lines of automated vehicles in disciplined lanes. Indian traffic mixes cars, buses, autos and two-wheelers with weak lane discipline and frequent cut-ins. Whether ACC vehicles damp or amplify stop-and-go waves in such traffic depends on their tuning, the human drivers around them and how often the line is broken by cut-ins. This matters for congestion, energy use and rear-end safety as ACC becomes common in Indian cars.

Objectives

• Model human drivers of different vehicle types with calibrated car-following models.
• Model ACC vehicles with realistic actuator lag, sensor delay and time-gap policies.
• Derive frequency-domain string stability conditions for mixed strings.
• Simulate long mixed strings with cut-ins and measure wave amplification.
• Propose and test ACC tunings that improve head-to-tail stability in mixed traffic.

What you need

Key concepts

Head-to-tail string stability

In a mixed string, individual vehicles may amplify disturbances, but the string as a whole is stable if the disturbance at the tail is no larger than at the head. It is the relevant measure when only some vehicles are automated.

Linearised car-following model

A non-linear model such as IDM can be linearised around an equilibrium speed and gap. The resulting transfer functions allow frequency-domain stability analysis.

Penetration rate

The share of vehicles in a string that use ACC. The effect of ACC on wave growth depends strongly on it and on where ACC vehicles are placed.

Cut-in disturbance

A vehicle entering a gap suddenly reduces the follower's gap, forcing it to brake. Frequent two-wheeler and car cut-ins are a major source of disturbance in Indian traffic.

Step-by-step guide

1. 1

   Frame the research questions

   State the questions: under what combinations of ACC time gap, controller gains, actuator lag, penetration rate and vehicle mix are mixed strings head-to-tail stable, and how do frequent cut-ins change this? Define metrics: peak frequency-domain gain, time-domain amplification of speed oscillations from head to tail, minimum TTC and rear-end conflicts. Agree a representative vehicle mix and speed range for Indian urban and highway traffic with your guide.

   Checkpoint: A research plan with questions, metrics, vehicle mix and speed ranges.

2. 2

   Model human drivers

   Use IDM or Gipps models for human drivers of cars, buses, autos and two-wheelers, with parameter sets for each vehicle type. Calibrate them from published Indian traffic studies where available, or from trajectory data you can legally use, and state the source of each parameter. Include reaction delay. Check each model reproduces plausible equilibrium gaps and acceleration at typical speeds.

   Checkpoint: Parameter tables for each vehicle type with sources and equilibrium checks.

3. 3

   Model ACC vehicles

   Model ACC vehicles with a constant time-gap policy, feedback on gap error and relative speed, a first-order actuator lag and a sensor delay. Use realistic values for production ACC and treat time gap and gains as design parameters. Validate the ACC model against standard behaviour, such as settling to the set time gap after a lead speed change, before using it in strings.

   Checkpoint: ACC step responses with correct steady time gap and damping for several gain sets.

4. 4

   Derive string stability conditions

   Linearise each model around equilibrium speeds and compute its transfer function from predecessor speed to own speed, including delays. Use Control System Toolbox to compute the peak gain for each vehicle type and ACC tuning. For a mixed string, multiply the transfer functions along the string and compute the head-to-tail peak gain. Map regions of stability against time gap, gains, lag and equilibrium speed.

   Checkpoint: Stability maps for individual vehicles and for representative mixed strings.

5. 5

   Build the time-domain simulation

   In Simulink, or Python for large runs, simulate strings of 20 to 40 vehicles with non-linear models and limits, placing ACC vehicles randomly at a chosen penetration rate. Apply head disturbances such as a braking pulse or a sine-wave speed variation. Measure speed oscillation amplitude at each position and the head-to-tail amplification ratio. Check that small-disturbance results agree with the linear analysis.

   Checkpoint: Simulated amplification ratios that match linear predictions for small disturbances.

6. 6

   Model cut-ins

   Add random cut-in events where a two-wheeler or car enters a gap with a chosen rate, gap position and relative speed, and cut-outs where a vehicle leaves. In Stateflow, model how ACC vehicles switch target and how human models react to the sudden gap change. Calibrate cut-in rates from published observations or justify ranges. Measure how cut-ins inject and amplify disturbances along the string.

   Checkpoint: Simulations where cut-ins produce visible waves, with amplification measured downstream.

7. 7

   Run the parameter study

   Sweep penetration rate (0 to 100 per cent), ACC time gap, controller gains, actuator lag, vehicle mix and cut-in rate. For each combination, run Monte Carlo trials with random ACC placement and cut-ins. Record head-to-tail amplification, minimum TTC, number of rear-end conflicts and average throughput. Store raw results so new metrics can be calculated later. Use a fractional design where the full grid is too large, and say so.

   Checkpoint: A complete results database covering every combination with Monte Carlo statistics.

8. 8

   Analyse the effect of ACC placement

   For a fixed penetration rate, compare random placement with ACC vehicles clustered together and evenly spread. Study whether an ACC vehicle directly behind a bus or a two-wheeler behaves differently. This placement analysis is important because penetration rate alone does not predict stability, and it suggests how deployment patterns affect traffic flow. Run enough random placements for the comparison to be statistically meaningful.

   Checkpoint: Amplification and safety results for clustered, spread and random placement.

9. 9

   Design improved ACC tunings

   Use the stability maps to propose ACC tunings that improve head-to-tail stability in mixed traffic, for example larger time gaps at certain speeds, adjusted relative speed gain or a filter that damps the frequencies human drivers amplify most. Check that comfort and cut-in response remain acceptable. Compare these tunings with a typical production-style tuning on the full parameter study. Keep the proposed tunings simple enough to implement in a production ACC.

   Checkpoint: A comparison of proposed and baseline tunings on stability, comfort and safety metrics.

10. 10

    Check robustness

    Test the proposed tunings against model uncertainty: different human driver parameters, longer actuator lags and sensor delay, and higher cut-in rates. Report where the proposed tuning keeps its advantage and where it fails. Check energy use with a simple EV power model as an additional outcome, since damping waves should also reduce energy waste in stop-and-go traffic. Report the spread of results across uncertain parameters, not only the mean.

    Checkpoint: Robustness results for the proposed tunings, including energy use.

11. 11

    Write up the research

    Present the vehicle models and their sources, the linear analysis, time-domain simulations, cut-in modelling, parameter study, placement analysis, proposed tunings and robustness. Distinguish clearly between analytical and simulated findings. Discuss limitations, including calibration data for Indian traffic, single-lane modelling and the absence of lateral behaviour. Release models and the results database. Summarise the recommended ACC tunings in a single table with the conditions under which each applies, so the findings are useful to practising engineers.

    Checkpoint: Thesis chapters with stability maps, simulation results and released models and data.

Testing & validation

• Human driver model parameters with sources and equilibrium checks.
• Linear stability predictions confirmed by small-disturbance simulations.
• Head-to-tail amplification across penetration rate, time gap, vehicle mix and cut-in rate.
• Placement effects on stability and safety.
• Proposed ACC tunings compared with a baseline tuning, including robustness and energy use.

Expected results

Stability maps and simulation evidence showing how ACC tuning, penetration rate, vehicle mix and cut-ins affect wave growth in mixed Indian traffic, with proposed ACC tunings that improve head-to-tail stability.

Take it further

• Add lateral behaviour and lane-less two-wheeler movement in a two-dimensional traffic simulator.
• Include CACC vehicles with V2V and study the effect of mixed ACC and CACC.
• Calibrate human models from trajectory data recorded with drones at an Indian junction, where permitted.
• Study the effect of ACC tuning on rear-end conflict rates with a surrogate safety measure.

Viva and review questions

Learn it at DIYguru

These DIYguru courses and programmes teach what this project needs.