Source-linked AI summary
Cooperative Adaptive Cruise Control for Connected Autonomous Vehicles by Factoring Communication-Related Constraints
Chaojie Wang, Siyuan Gong, Anye Zhou, Tao Li, Srinivas Peeta
TL;DR
CACC research often assumes fixed communication topologies despite V2V failures that dynamically change platoon information flow. This paper introduces CACC-OIFT, which optimizes the topology and adapts PD control to unfolding failures; it reports improved string stability in unreliable V2V conditions and identifies scope extensions for future work.
Problem
Existing CACC designs commonly assume fixed IFTs, while communication failures and congestion can dynamically change information flow and require control across degeneration scenarios.
Method
CACC-OIFT proactively activates or deactivates V2V send functions to optimize expected string stability, then uses an adaptive PD controller for each unfolding degeneration scenario.
Results
The study reports that rigorous proactive use of IFT dynamics improves platoon performance and that CACC-OIFT enhances string stability in unreliable V2V communication contexts.
Takeaways & Limitations
Optimizing communication topology together with adaptive control provides a practical strategy for maintaining string stability while accounting for communication failures.
Takeaways & Limitations
The study does not consider receiver failures, actuator delay, nonlinear vehicle dynamics, external disturbances, or mixed traffic flow.
Abstract
from arXiv · showhide
Emergent cooperative adaptive cruise control (CACC) strategies being proposed in the literature for platoon formation in the Connected Autonomous Vehicle (CAV) context mostly assume idealized fixed information flow topologies (IFTs) for the platoon, implying guaranteed vehicle-to-vehicle (V2V) communications for the IFT assumed. Since CACC strategies entail continuous information broadcasting, communication failures can occur in congested CAV traffic networks, leading to a platoon's IFT varying dynamically. To enhance the performance of CACC strategies, this study proposes the idea of dynamically optimizing the IFT for CACC, labeled the CACC-OIFT strategy. Under CACC-OIFT, the vehicles in the platoon cooperatively determine in real-time which vehicles will dynamically deactivate or activate the "send" functionality of their V2V communication devices to generate IFTs that optimize the platoon performance in terms of string stability under the ambient traffic conditions. Given the adaptive Proportional-Derivative (PD) controller with a two-predecessor-following scheme, and the ambient traffic conditions and the platoon size just before the start of a time period, the IFT optimization model determines the optimal IFT that maximizes the expected string stability. The optimal IFT is deployed for that time period, and the adaptive PD controller continuously determines the car-following behaviors of the vehicles based on the unfolding degeneration scenario for each time instant within that period. The effectiveness of the proposed CACC-OIFT is validated through numerical experiments in NS-3 based on NGSIM field data. The results indicate that the proposed CACC-OIFT can significantly enhance the string stability of platoon control in an unreliable V2V communication context, outperforming CACCs with fixed IFTs or with passive adaptive schemes for IFT dynamics.
1. Introduction
Existing CACC research commonly assumes fixed, idealized information flow topologies, although V2V failures can dynamically alter platoon communication. CACC-OIFT proactively optimizes the topology and adapts control to degeneration scenarios to improve expected string stability.
- Research gap: Most CACC studies assume identical, predetermined VIFTs and a fixed platoon-level IFT, overlooking dynamic changes caused by V2V failures.Failures may arise from communication interference or information congestion, particularly in congested CAV traffic.
- Research gap: Fixed-IFT CACC may execute erroneous control actions or degrade to ACC when communication failures occur, reducing mobility, stability, and safety performance.
- Proposed strategy: CACC-OIFT proactively activates or deactivates vehicles’ V2V “send” functions to select an IFT that maximizes expected string stability under ambient conditions.The strategy combines an IFT optimization model with an adaptive PD controller.
- Operational deployment: The optimal IFT is computed before each time period, deployed throughout that period, and paired with controller behavior that responds to the degeneration scenario unfolding at each instant.The model assumes ambient traffic conditions and platoon size remain unchanged within the time period.
- Evaluation: The study develops a two-step solution algorithm, proves structural properties, and evaluates CACC-OIFT with NGSIM data in the NS-3 network simulator.The reported validation concerns unreliable V2V communication and compares against fixed-IFT and passive adaptive approaches.
- Proposed strategy: Expected string stability is evaluated by summing frequency-domain speed-oscillation energy across degeneration scenarios, with lower energy indicating better damping.The expected oscillation energy is a probability-weighted sum over possible degeneration scenarios.
2. IFT and degeneration scenarios
The study represents a platoon’s information flow topology with binary communication-status variables and models its possible sender-failure degeneration scenarios. Vehicles then select among communication-dependent controller statuses based on the information successfully received.
- IFT representation: An IFT for N+1 vehicles is encoded as ξ=[η_0,η_1,…,η_N], where η_i=1 means vehicle i can send and η_i=0 means its “send” functionality is deactivated.The set Ω contains all IFTs consistent with the two-predecessor-following scheme.
- Degeneration scenarios: The figures distinguish an initial IFT from degeneration cases in which CAV 2, CAV 0, or both fail to send messages.
- Degeneration scenarios: An IFT can degenerate when one or more activated senders fail, producing configurations obtained by changing activated sender statuses from 1 to 0.For ξ=[1,0,1,0,0], four degeneration scenarios are possible, including simultaneous failures of vehicles 0 and 2.
- Receiver-side control: A following vehicle uses CACC1 when both predecessors transmit successfully; one failed sender produces alternative CACC statuses based on the information still received.The receiver-side communication status determines the applicable car-following controller.
3. Formulation of optimization model for the IFT
The IFT optimization model evaluates candidate communication topologies by combining degeneration-scenario probabilities with platoon speed-oscillation energy, using the resulting expectation as a string-stability objective. It incorporates sender-failure probabilities determined by ambient traffic and communication contention.
- Optimization model: The optimization objective selects the IFT with the lowest expected speed-oscillation energy, corresponding to higher expected string stability.Expected performance is computed across all degeneration scenarios and their probabilities.
- Optimization model: The model represents each IFT as a binary send-activation vector and enumerates feasible degeneration scenarios by allowing activated senders to fail.Scenario probabilities must sum to one for every candidate IFT.
- Speed oscillation energy: Speed-oscillation energy treats vehicle speed oscillations as frequency-domain signals and sums energy across frequencies and platoon vehicles.The vehicle-level energy is based on the squared frequency-domain speed amplitude.
- Speed oscillation energy: The model links each following vehicle’s oscillation recursively to predecessor oscillations and the characteristics of the CACC controller.The leading vehicle trajectory is assumed to follow the ambient traffic oscillation.
- String-stability linkage: For a fixed leading trajectory, lower platoon speed-oscillation energy implies a lower transfer-function magnitude and better string-stability performance.The transfer function is evaluated separately for controller configurations and degeneration scenarios.
- Degeneration probabilities: Sender failures are modeled as broadcast collisions among activated vehicles, while larger contention windows or retransmissions reduce collisions at the cost of information delay.The study focuses on sending-process failures rather than receiver-related failures.
- Degeneration probabilities: A Markov-chain contention model estimates each sender’s success probability from channel busyness, activated vehicles within communication range, and ambient traffic density.The resulting sender success rates are used to derive degeneration-scenario probabilities for each IFT.
4. Formulation of adaptive controller for degeneration scenarios of IFT
The adaptive controller combines spacing-error feedback with predecessor-acceleration feedforward and adjusts these components to the active IFT and its degeneration scenarios. Stability analysis uses the trajectory transfer function to select controller parameters under conservative conditions.
- Controller formulation: The proposed controller uses a two-predecessor-following scheme and adapts its feedback and feedforward structure to each IFT degeneration scenario.The transfer function SS_X,i is used as the string-stability indicator in the optimization model.
- Control structure: The control command sums a spacing-error feedback term with feedforward terms based on acceleration rates from the immediate and second predecessors.The second vehicle receives feedforward acceleration information only from the leading vehicle.
- Control structure: The ideal vehicle model treats acceleration as input and position as output, with longitudinal dynamics represented by G_i(s)=s^-2.The state variables are position, velocity, and acceleration.
- Control structure: The constant-time-headway spacing policy sets desired distances to one or two predecessors as L+h ẋ_i and twice that quantity.L is standstill distance and h is the desired time headway.
- Control structure: The feedback loop uses weighted spacing errors from both predecessors, with weights determined by the vehicle’s controller status.The weights sum to one across the controller statuses.
- Control structure: Feedforward filters are designed under a zero-error condition to eliminate spacing error between adjacent vehicles.The filters process predecessor acceleration information before it contributes to the control command.
- Parameter determination: The PD cutoff frequency affects the string-stability transfer function, while time headway and cutoff frequency are analyzed jointly for noise mitigation and stability.The deployed controller uses a discretized version of the control command.
- Stability analysis: Under the conservative marginal-stability case, the derived feasible region requires positive time headway for CACC1, CACC2, and CACC3.The analysis imposes a worst-case condition in which predecessor-to-leader trajectory ratios equal one.
5. Algorithm to solve OPT-II
The algorithm solves OPT-II numerically because binary IFT decisions, degeneration-specific transfer functions, and frequency-domain traffic oscillations make the objective complex. It exploits controller properties and repeated degeneration scenarios to improve efficiency, then evaluates communication and control performance under realistic traffic conditions.
- Computational challenge: OPT-II is a discrete optimization problem whose objective varies across degeneration scenarios and requires frequency-domain ambient-traffic oscillations.The candidate IFT count and scenario count grow with platoon size, increasing computational burden.
- IFT structure: The last vehicle should deactivate its “send” functionality because it only receives information, reducing collision probability and improving V2V reliability.
- Controller comparison: CACC transfer functions damp oscillations better than ACC, with |SS_X,i,1| < |SS_X,i,2| < |SS_X,i,4| and |SS_X,i,1| < |SS_X,i,3| < |SS_X,i,4|.The comparison is based on cut-off frequencies, where smaller values indicate better string stability.
- IFT structure: The leading vehicle should activate its “send” functionality in the optimal IFT because shifting activation toward it produces lower expected oscillation energy.
- Computational efficiency: Identical degeneration scenarios across IFTs allow the fully activated pattern except for the last vehicle, ξ = [1, …, 1,0], to represent scenarios for other IFTs.This reduces repeated string-stability calculations during optimization.
- Communication performance: Higher traffic density or a larger share of activated “send” functionalities lowers communication success rates by increasing contention and information collisions.The communication experiments use average densities from 25 to 40 vehicles/km.
- Optimization results: Optimal IFTs balance more CACC operation against collision risk by placing deactivated “send” functionalities after activated groups, especially at higher density.For k̅=25 and N=14, the pattern 111000 appears in both the first and last six vehicles.
- Control performance: CACC-OIFT reduces spacing and speed fluctuations more than CACC-DIFT and CACC-FIFT in the reported platoon experiments.For i=2, maximum spacing error is 1.05m under CACC-OIFT, compared with 1.42m under CACC-DIFT and 1.51m under CACC-FIFT.
7. Concluding comments
CACC-OIFT dynamically optimizes platoon information flow topology under communication failures, using an adaptive PD controller and expected string-stability optimization. The study reports more reliable information flow, better damping and stabilization, while identifying extensions beyond the pure-CAV setting.
- CACC-OIFT dynamically activates or deactivates vehicle communication devices to optimize expected string stability under ambient traffic conditions.The model considers all degeneration scenarios and deploys the selected topology with an adaptive PD controller.
- The approach explicitly incorporates information-flow dynamics and communication failures into CACC design for pure CAV platoons.The authors describe this as the first attempt to leverage information-flow dynamics for CACC performance enhancement.
- CACC-OIFT generates more reliable information flow, dampens traffic-oscillation propagation, and stabilizes traffic flow across the platoon.
- CACC-OIFT is reported as string stable and as considerably outperforming CACC-DIFT and CACC-FIFT.
- Future extensions include receiver failures, richer vehicle and actuator models, alternative controllers, and mixed-flow CACC designs.The current study assumes a pure CAV platoon and identifies mixed-flow operation as a future direction.