An Online Evolving Framework for Modeling the Safe Autonomous Vehicle Control System via Online Recognition of Latent Risks

Dimitar Filev; Teawon Han; Umit Ozguner

REVIEW 3 major objections 7 minor 18 references

An Online Evolving Framework for Modeling the Safe Autonomous Vehicle Control System via Online Recognition of Latent Risks

T0 review · 3 major / 7 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read An online-evolving finite state machine can learn a collision-prone Dead-End state from scratch and predict one-step-ahead driving states accurately in simulations.

desk verdict Modest extension of the authors' own prior work, with a real reproducibility gap (undefined variance in Eq. 7) that undercuts the headline claims, but the core idea is worth refereeing. read the letter →

arxiv 1908.10823 v1 pith:TLQTQ2HJ submitted 2019-08-28 cs.RO cs.LGcs.NE

Teawon Han , Dimitar Filev , Umit Ozguner This is my paper · ORCID

classification cs.ROcs.LGcs.NE

keywords evolvingfinitestatemachineonlineclusteringlatentriskdetectionautomatedvehiclesafetyMarkovchaincar-followingsimulationJensen-ShannondivergenceDead-End

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper proposes the evolving Finite State Machine (e-FSM), a model that builds its own set of driving states online from sensor observations and learns how those states transition under each available action. The authors aim to show that this lets an automated vehicle controller recognize situations it was never pre-programmed to expect, including a collision-prone 'Dead-End' state, and choose safer actions. In repeated car-following simulations with different driver styles, they report that every collision is recognized as the same state (state #3), and that one-step-ahead state probabilities predicted from the learned transition matrices match the observed recognition within a Jensen-Shannon divergence below 0.15. If the claim holds, an AV controller could detect latent risks online and compare candidate actions by their predicted future states, without a hand-designed list of dangerous situations.

What carries the argument

The load-bearing object is the evolving Finite State Machine (e-FSM): a growing set of states, each represented by the center of a cluster of observations, plus one transition matrix for each discrete action. A new state is created when the evolving Takagi-Sugeno online clustering rule decides the current observation has higher potential than all existing cluster centers and is farther than threshold epsilon from the nearest center; otherwise the observation is assigned to clusters by a similarity function that produces a probability distribution over states. Each transition matrix is updated by a stochastic approximation rule using the previous and current state probability distributions, and its dimension grows whenever a state is added. The mechanism that carries the argument is the combination of these two devices: clustering turns continuous observations into discrete states online, and action-conditioned Markov matrices make one-step-ahead prediction a simple matrix-vector product.

What would settle it

Repeat the same car-following experiment with different eTS thresholds (for example epsilon=0.1 or epsilon=0.5) or with a differently scaled observation vector; if collision events then map to more than one state index, or the Jensen-Shannon divergence between predicted and recognized states exceeds 0.15 repeatedly, the paper's uniqueness and precision claims would be contradicted.

Watch

Extended reading notes

Core claim

The central claim is that an online-evolving state machine can determine, recognize, and predict driving-risk states without a predefined state set. States are cluster centers of observations (headway, follower speed, preceding-vehicle speed) created by an online clustering rule; recognition is a probability distribution over these clusters; and state transitions are stored in one Markov transition matrix per discrete action, updated online and expanded whenever a new state appears. Using an Intelligent Driver Model controller in a simulated car-following scenario with four controller configurations repeated 20 times, the paper reports that a Dead-End state—the situation in which a collision is no longer avoidable—is always recognized as state #3, and the Jensen-Shannon divergence between predicted and recognized one-step-ahead state distributions is less than 0.15 except for the very first prediction. The paper concludes that the Dead-End state is uniquely determined and consistently recognized and that the identified transition matrices are accurate enough for future-state prediction.

Load-bearing premise

The load-bearing premise is that the online clustering rule, with its arbitrarily chosen thresholds epsilon=0.3 and rho=0.85 and an unspecified cluster variance, partitions observations so that each cluster corresponds to one unique driving situation; the claim that every Dead-End event maps to state #3 depends entirely on that partition.

Editorial extensions

If this is right

An AV controller can be given an early warning whenever the current observation matches the learned Dead-End state, before the collision actually occurs.
Because each action has its own transition matrix, the controller can compare candidate actions by the future-state distributions they imply, rather than only by immediate reward.
The state set is not fixed in advance: the model can keep adding states as new situations appear, and in the reported experiments the number of states stabilizes at 7 after four simulations.
The same recognized state index can serve as a shared risk label across different driver styles and controller settings, since the identical unsafe situation maps to state #3 in all cases.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The paper does not specify how the cluster variance in the similarity function is computed; a natural extension would be to test whether the uniqueness of the Dead-End state survives alternative variance definitions or normalized observations.
The stability of the result under different eTS thresholds is untested; changing epsilon or rho could produce a different partition and possibly split or merge the collision state, so a threshold sensitivity study would delimit the claim.
The observation vector contains only longitudinal quantities; extending e-FSM to lateral position, lane context, or map information would test whether the same unique-state property holds in more realistic traffic situations.
If the approach scales, the same mechanism could label risky states in other safety-critical systems, such as robotic manipulation or human-robot interaction, where unanticipated states must be recognized online rather than enumerated.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

You should know this before reading: the paper is a modest extension of the authors' own prior work (Filev-Kolmanovsky generalized Markov models plus eTS clustering), with one genuinely new twist: action-conditioned transition matrices. The application to latent-risk state recognition in automated vehicles is fresh, but the machinery underneath is established. If you are expecting a fundamentally new method, you will be disappointed; if you are open to an incremental but plausible combination, this is a reasonable read.

What the paper does well: it lays out the e-FSM framework clearly, explains how states are determined online via eTS and recognized via probability distributions, and demonstrates, in a simple car-following setup, that a collision-prone Dead-End state is consistently mapped to a single cluster (state #3). The JSD values below 0.15 between predicted and recognized state distributions are a positive signal, though not a strong one. The writing is straightforward and the experimental setup is easy to follow.

Now the soft spots, in proportion. The stress-test note is correct: Eq. 7 uses var(z*_i) without defining it. A cluster center is a vector, so its variance is not well-defined without an explicit convention (e.g., variance of the observations assigned to that cluster). Since every probability distribution in the paper flows through this equation, the experimental results are not reproducible as written. That is a real blocker, not a nit. Second, the 20 repeated runs per case appear to be deterministic IDM simulations with no stochastic perturbation, so "without exception" may just mean the same trajectory was replayed 20 times. That weakens the consistency claim considerably. Third, the transition-prediction evaluation is in-sample: the same recognized state sequence used to update the matrices is used to test the one-step-ahead predictions, so low JSD can reflect fitting rather than predictive accuracy. Fourth, no sensitivity analysis is reported for the arbitrary parameters (epsilon, rho, phi, delta), and the safety framework's "action reviser" is only sketched, not implemented. The claim about "choosing a better action" is therefore not actually demonstrated.

Overall, the central idea is sound: online evolving state recognition coupled with action-conditioned Markov transitions is a plausible building block for AV decision-making. The paper is not a breakthrough, but it is a coherent, honest piece of work with one serious omission (the undefined variance) and several addressable weaknesses. A serious editor should send it to peer review, expecting major revision. I would not cite it in my own work yet, but I would bring it to a reading group to discuss what needs to be fixed before the claims can be trusted.

Referee Report

3 major / 7 minor

Summary. The paper proposes an online evolving Finite State Machine (e-FSM) for automated vehicle control. The e-FSM uses evolving Takagi-Sugeno (eTS) online clustering to determine states uniquely as new observations arrive, computes a probability distribution over recognized states via a similarity function, and learns action-conditioned transition matrices using the online stochastic method of Filev and Kolmanovsky. The framework is illustrated in a SUMO car-following scenario with IDM controllers of different aggressiveness, run 20 times per case for four controller-switch cases. The authors claim that a collision-prone Dead-End state is uniquely determined and consistently recognized as state #3, and that one-step-ahead state predictions, validated by Jensen-Shannon divergence (JSD) below 0.15, show the transition matrices are precisely identified.

Significance. If the claims hold, the e-FSM would be a useful online component for detecting latent driving risks and predicting future states, with a concrete mechanism for evolving the state set without a fixed reward function or pre-enumerated scenarios. The paper provides explicit equations for clustering, recognition, and transition learning, and it makes a falsifiable prediction about the repeatability of state identification. However, two load-bearing issues currently prevent the claims from being accepted as demonstrated: the recognition probabilities depend on an undefined quantity in the similarity function, and the transition-validation protocol is in-sample. The paper's strengths are its clear formulation of the online learning problem and its explicit experimental setup, but the evidence as presented is not yet reproducible.

major comments (3)

[§2.2.2, Eq. (7)] The similarity function η_i^t(z_t) = exp( -(z_t - z*_i)^T (z_t - z*_i) / var(z*_i) ) uses the quantity var(z*_i), the variance of a cluster center, which is never defined anywhere in the manuscript. A cluster center is a single vector, so its variance is not well-defined without an explicit convention such as the variance of the observations assigned to that cluster, a per-dimension variance, or a user-specified scalar. Because Prob(S_t) is computed from this equation, and because the transition-matrix updates (Eqs. 15–17), the predicted distributions (Eq. 18), and the JSD results in Figure 5 all depend on Prob(S_t), the experimental evidence cannot be independently reproduced. The claim that the Dead-End state is uniquely and consistently recognized as state #3 may be an artifact of an unreported implementation choice for this variance. The authors should define var(z*_i) precisely, state how it is computed, and report the resulting state-recognition results.
[§3.2.2, Figure 5] The validation of the transition matrices is entirely in-sample. The transition matrix Pa(r)_t is updated online using τ(t) = Prob(S_{t-1}) and γ(t) = Prob(S_t) from the same simulation run, and then Prob_pred(S_{t+1}) = Pa(r)_t · Prob(S_t) is compared with the same run's Prob(S_{t+1}). Low JSD therefore measures how well the model fits the exact data used for its own updates, not how precisely the state-transitions are identified in any predictive sense. To support the claim that the transition matrices are precisely identified, the authors should perform out-of-sample validation, for example by training on the first portion of each trajectory and testing on held-out later portions, by cross-validating across the 20 runs, or by comparing the JSD against a baseline model such as the empirical marginal transition frequencies.
[§3.2, experiment process] The paper states that each of the four cases is simulated 20 times, giving 80 simulations in total, but no source of stochasticity is specified anywhere in the experimental setting. The SUMO and IDM models are deterministic given the initial conditions and parameters, so it is likely that the 20 repetitions are identical replays of the same trajectory. If so, the statement that state #3 is recognized 'without exception' in all 80 simulations only demonstrates self-consistency of a single deterministic run, not consistency across varied conditions. The authors should either introduce explicit stochastic perturbations (e.g., in initial speeds, IDM parameters, or observation noise) and report the distribution of recognized states across runs, or explicitly state that the repetitions are identical and temper the consistency claim accordingly.

minor comments (7)

[§2.2.1, Eq. (4)] The initialization of the potential recursion is incomplete: the text says b1 is initialized to 0 and the first input is set as a cluster center with potential 1, but Equation (4) defines b_t = b_{t-1} + z_{t-1}^T z_{t-1}, which needs a specification for t=1 and a consistent treatment of z_0.
[§2.2.2, Eq. (7)] The notation for cluster centers is inconsistent across equations: Equation (6) uses z*_t^i, while Equation (7) uses z*_i and later text uses z*_i^t. The authors should standardize the superscript/subscript convention to avoid ambiguity.
[§3.1, experimental settings] The learning rate φ in Equations (13)–(17) is never given a numerical value. Since the transition-matrix identification results depend on this rate, the authors should report the value used in the simulations.
[§3.1, experimental settings] The eTS parameters ε=0.3 and ρ=0.85 are described as arbitrarily assigned, which is acceptable for a demonstrative study, but the paper should include a sensitivity analysis or at least a discussion of how the number of states and the recognition of state #3 depend on these thresholds.
[§3.1, action encoding] The action discretization description appears inconsistent: the continuous action set Ac = [−2.5, 2.5] m/s^2 is said to be encoded with a range of 0.3 m/s^2 into 17 intervals, but the arithmetic (5.0/0.3 ≈ 16.67) does not produce exactly 17 intervals. The exact bin edges and whether the endpoints are inclusive should be clarified.
[§3.2.2, Figure 5] The claim that JSD values below 0.15 indicate accurate prediction would be strengthened by a stated threshold or a comparison with a null distribution; without such context, 'significant similarity' is asserted rather than demonstrated.
[§4, framework overview] The action reviser module is introduced but its operation is not described; the paper focuses on e-FSM, so this is acceptable as a future-work direction, but the authors should mark it explicitly as outside the scope of the current validation.

Circularity Check

1 steps flagged · score 3.0 of 10

The transition-matrix validation is an in-sample fit of the same recognition probabilities, but the DE-state claim is anchored to an external collision label; no definitional or self-citation circularity is present.

fitted input called prediction [Section 2.3.2 (Eqs. 15-17) and Section 3.2.2 (Eq. 18)]
"Because multiple transition matrices are implemented in e-FSM for representing state-transitions based on the chosen actions, Equation 12, 13, and 14 are re-defined by Equation 15, 16, and 17, where r ={1,2,...,q}, τ(t) = Prob(St−1), and γ(t) = Prob(St). ... To show how accurate the transition matrices are identified through proposed methods, the probability distributions of predicted and recognized state, Probpred(St+1) and Prob(St+1), are compared. ... Probpred(St+1) = Pa(r)t · Prob(St)"

The transition matrix Pa(r)t is fitted online from the same recognized-state probabilities that are later used as the input to the 'prediction': Eq. 16 updates Fa(r)(t) with τ(t)γ(t)T, where γ(t)=Prob(St), and Eq. 18 multiplies that fitted matrix by the same style of recognized distribution to obtain Probpred(St+1). The JSD comparison in Section 3.2.2 therefore measures how well the Markov model reproduces the very recognition sequence used to estimate it, i.e., an in-sample goodness-of-fit. The paper then concludes that 'state-transitions are precisely identified,' but the low JSD is not independent evidence; it is a self-consistency check of the fitted model.

full rationale

The central DE-state claim is not circular: the identification of state #3 with the Dead-End state is validated against an external event (collision) observed in the SUMO/IDM simulation, not derived from the e-FSM equations themselves. The state-recognition and transition-identification method is also presented with explicit equations and builds on published clustering and Markov-model work, so there is no hidden uniqueness theorem or ansatz smuggled in via self-citation. The only circularity-like element is the validation of the transition matrices: the matrices are fitted from Prob(St) and then used to predict Prob(St+1), which is compared with the recognized Prob(St+1) from the same model. This is partial in-sample fitting rather than a definitional equivalence, because at the moment of prediction the future recognized distribution has not yet been inserted into the transition update. I therefore assign a score of 3, not higher. Separately, Eq. 7 contains the undefined term var(z*_i), the variance of a cluster center, which is a serious reproducibility defect and could affect the uniqueness claim, but it is a correctness concern, not a circularity concern under the stated rules.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central claim relies on several arbitrary or unreported settings: the eTS thresholds ε and ρ, the learning rate φ, the initialization ε̄, and the undefined variance in the recognition similarity. These are not derived from data or theory, and the paper gives no sensitivity analysis. The Markov and sufficiency assumptions are also load-bearing for the state-recognition and prediction claims.

free parameters (6)

ε (eTS distance threshold) = 0.3
Set in Section 3.1; controls how often new clusters are created, directly affecting the number of states and the uniqueness of the Dead-End state.
ρ (eTS potential decay factor) = 0.85
Set in Section 3.1; used in Eq 6 to update cluster potentials; affects the dynamics of state creation.
φ (transition matrix learning rate) = not reported
Used in Eq 16-17 for updating F matrices; the paper does not provide its value, yet the convergence and accuracy of the identified transitions depend on it.
ε̄ (initialization constant for transition matrices) = not reported
Used in Section 2.3.2 to initialize F and F_o for new states; value not specified.
δ (action discretization interval) = 0.3 m/s²
Used in Section 3.1 to encode continuous acceleration into 17 discrete actions; affects the granularity of action-conditioned transition matrices.
var(z*_i) (cluster variance in similarity function) = undefined
Appears in Eq 7 for state recognition; no definition or estimation method is given, making the similarity values ill-posed.

assumptions (5)

domain assumption Markov property: future state depends only on current state and action
Assumed in Eq 18-20 and implicit throughout; the car-following process is treated as a Markov chain over the learned clusters.
domain assumption Observation vector z_t is a sufficient state representation
The choice z_t = [headway, v_f, v_p] in Section 3.1 assumes these three variables capture all information relevant to collision risk.
ad hoc to paper eTS clustering thresholds ε and ρ are appropriate for the domain
The authors state ε and ρ are arbitrarily assigned (Section 2.2.1); no sensitivity analysis is provided.
ad hoc to paper Cluster centers have a well-defined variance so Eq 7 is computable
The paper never defines var(z*_i); without it, the recognition probabilities are undefined.
domain assumption The recursive transition update (Eq 16-17) converges to the true conditional probabilities
Inherited from ref 13 without proof or adaptation analysis for the action-conditioned case.

how reviews work

0 comments

Cite this review

Pith. "Pith review of An Online Evolving Framework for Modeling the Safe Autonomous Vehicle Control System via Online Recognition of Latent Risks." pith.science (2026). https://pith.science/paper/TLQTQ2HJ

@misc{pith2026190810823,
  author       = {Pith},
  title        = {Pith review of: An Online Evolving Framework for Modeling the Safe Autonomous Vehicle Control System via Online Recognition of Latent Risks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TLQTQ2HJ}},
  note         = {Machine review of arXiv:1908.10823}
}

read the original abstract

An online evolving framework is proposed to support modeling the safe Automated Vehicle (AV) control system by making the controller able to recognize unexpected situations and react appropriately by choosing a better action. Within the framework, the evolving Finite State Machine (e-FSM), which is an online model able to (1) determine states uniquely as needed, (2) recognize states, and (3) identify state-transitions, is introduced. In this study, the e-FSM's capabilities are explained and illustrated by simulating a simple car-following scenario. As a vehicle controller, the Intelligent Driver Model (IDM) is implemented, and different sets of IDM parameters are assigned to the following vehicle for simulating various situations (including the collision). While simulating the car-following scenario, e-FSM recognizes and determines the states and identifies the transition matrices by suggested methods. To verify if e-FSM can recognize and determine states uniquely, we analyze whether the same state is recognized under the identical situation. The difference between probability distributions of predicted and recognized states is measured by the Jensen-Shannon divergence (JSD) method to validate the accuracy of identified transition-matrices. As shown in the results, the Dead-End state which has latent-risk of the collision is uniquely determined and consistently recognized. Also, the probability distributions of the predicted state are significantly similar to the recognized state, declaring that the state-transitions are precisely identified.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 17 canonical work pages

[1]

Kurt, A. and Ü. Özgüner, Hybrid state system development for autonomous vehicle control in urban scenarios. IFAC Proceedings Volumes, V ol. 41, No. 2, 2008, pp. 9540–9545

work page 2008
[2]

Redmill, K. A., U. Ozguner, S. Biddlestone, A. Hsieh, and J. Martin, Ohio state university experiences at the darpa challenges. SAE International Journal of Commercial Vehicles , V ol. 1, No. 2008-01-2718, 2008, pp. 527–533

work page 2008
[3]

Liu, Y . and U. Ozguner, Human driver model and driver decision making for intersection driving. In 2007 IEEE Intelligent Vehicles Symposium, IEEE, 2007, pp. 642–647

work page 2007
[4]

Krishnamurthy, and U

Gadepally, V ., A. Krishnamurthy, and U. Ozguner, A framework for estimating driver deci- sions near intersections. IEEE Transactions on Intelligent Transportation Systems, V ol. 15, No. 2, 2013, pp. 637–646

work page 2013
[5]

Kurt, A. and Ü. Özgüner, A probabilistic model of a set of driving decisions. In 2011 14th International IEEE conference on intelligent transportation systems (ITSC) , IEEE, 2011, pp. 570–575

work page 2011
[6]

Noh, S. and K. An, Decision-making framework for automated driving in highway envi- ronments. IEEE Transactions on Intelligent Transportation Systems, V ol. 19, No. 1, 2017, pp. 58–71

work page 2017
[7]

Zhang, M., N. Li, A. Girard, and I. Kolmanovsky, A ﬁnite state machine based auto- mated driving controller and its stochastic optimization. In ASME 2017 Dynamic Sys- tems and Control Conference , American Society of Mechanical Engineers, 2017, pp. V002T07A002–V002T07A002

work page 2017
[8]

Barjasteh, H

Al-Qizwini, M., I. Barjasteh, H. Al-Qassab, and H. Radha, Deep learning algorithm for autonomous driving using googlenet. In 2017 IEEE Intelligent Vehicles Symposium (IV) , IEEE, 2017, pp. 89–96

work page 2017

Show all 18 references

[9]

Del Testa, D

Bojarski, M., D. Del Testa, D. Dworakowski, B. Firner, B. Flepp, P. Goyal, L. D. Jackel, M. Monfort, U. Muller, J. Zhang, et al., End to end learning for self-driving cars. arXiv preprint arXiv:1604.07316, 2016

2016 arXiv
[10]

Wolff, and L

Hoel, C.-J., K. Wolff, and L. Laine, Automated speed and lane change decision making using deep reinforcement learning. In 2018 21st International Conference on Intelligent Transportation Systems (ITSC), IEEE, 2018, pp. 2148–2155

2018
[11]

You, C., J. Lu, D. Filev, and P. Tsiotras, Highway trafﬁc modeling and decision making for autonomous vehicle using reinforcement learning. In 2018 IEEE Intelligent Vehicles Symposium (IV), IEEE, 2018, pp. 1227–1232

2018
[12]

Hejase, M., A. Kurt, T. Aldemir, and U. Ozguner, Identiﬁcation of Risk Signiﬁcant Auto- motive Scenarios Under Hardware Failures. arXiv preprint arXiv:1804.04348, 2018

2018 arXiv
[13]

Filev, D. P. and I. Kolmanovsky, Generalized markov models for real-time modeling of continuous systems. IEEE Transactions on Fuzzy Systems, V ol. 22, No. 4, 2013, pp. 983– 998

2013
[14]

Angelov, P. P. and D. P. Filev, An approach to online identiﬁcation of Takagi-Sugeno fuzzy models. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), V ol. 34, No. 1, 2004, pp. 484–498

2004
[15]

Filev, D., J. Lu, F. Tseng, and K. Prakah-Asante, Real-time driver characterization during car following using stochastic evolving models. In 2011 IEEE International Conference on Systems, Man, and Cybernetics, IEEE, 2011, pp. 1031–1036. Han, Filev, and Özgüner 16

2011
[16]

Treiber, and D

Kesting, A., M. Treiber, and D. Helbing, Enhanced intelligent driver model to access the impact of driving strategies on trafﬁc capacity. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , V ol. 368, No. 1928, 2010, pp. 4585–4605

1928
[17]

Hennecke, and D

Treiber, M., A. Hennecke, and D. Helbing, Congested trafﬁc states in empirical observa- tions and microscopic simulations. Physical review E, V ol. 62, No. 2, 2000, p. 1805

2000
[18]

IEEE Transactions on Infor- mation theory, V ol

Lin, J., Divergence measures based on the Shannon entropy. IEEE Transactions on Infor- mation theory, V ol. 37, No. 1, 1991, pp. 145–151

1991

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Kurt, A. and Ü. Özgüner, Hybrid state system development for autonomous vehicle control in urban scenarios. IFAC Proceedings Volumes, V ol. 41, No. 2, 2008, pp. 9540–9545

work page 2008

[2] [2]

Redmill, K. A., U. Ozguner, S. Biddlestone, A. Hsieh, and J. Martin, Ohio state university experiences at the darpa challenges. SAE International Journal of Commercial Vehicles , V ol. 1, No. 2008-01-2718, 2008, pp. 527–533

work page 2008

[3] [3]

Liu, Y . and U. Ozguner, Human driver model and driver decision making for intersection driving. In 2007 IEEE Intelligent Vehicles Symposium, IEEE, 2007, pp. 642–647

work page 2007

[4] [4]

Krishnamurthy, and U

Gadepally, V ., A. Krishnamurthy, and U. Ozguner, A framework for estimating driver deci- sions near intersections. IEEE Transactions on Intelligent Transportation Systems, V ol. 15, No. 2, 2013, pp. 637–646

work page 2013

[5] [5]

Kurt, A. and Ü. Özgüner, A probabilistic model of a set of driving decisions. In 2011 14th International IEEE conference on intelligent transportation systems (ITSC) , IEEE, 2011, pp. 570–575

work page 2011

[6] [6]

Noh, S. and K. An, Decision-making framework for automated driving in highway envi- ronments. IEEE Transactions on Intelligent Transportation Systems, V ol. 19, No. 1, 2017, pp. 58–71

work page 2017

[7] [7]

Zhang, M., N. Li, A. Girard, and I. Kolmanovsky, A ﬁnite state machine based auto- mated driving controller and its stochastic optimization. In ASME 2017 Dynamic Sys- tems and Control Conference , American Society of Mechanical Engineers, 2017, pp. V002T07A002–V002T07A002

work page 2017

[8] [8]

Barjasteh, H

Al-Qizwini, M., I. Barjasteh, H. Al-Qassab, and H. Radha, Deep learning algorithm for autonomous driving using googlenet. In 2017 IEEE Intelligent Vehicles Symposium (IV) , IEEE, 2017, pp. 89–96

work page 2017