Debraj Chakraborty
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Debraj Chakraborty has a total of 33 co-authors. Here are the top co-authors by number of shared publications:
- Ameer Kalandar Mulla
Department of Electrical, Electronics and Communication Engineering, Indian Institute of Technology Dharwad, Dharwad, India
4 shared publications - Aditya Chaudhari
University of Florida Department of Mechanical Engineering Gainesville, FL 32601
3 shared publications - D. Manjunath
Department of Electrical Engineering, IIT Bombay, Mumbai, India
3 shared publications - Apurva Joshi
IITB-Monash Research Academy, Mumbai, India
2 shared publications - Rakesh Chavan
Electrical, Indian Institute of Technology Bombay, mumbai India 400092
2 shared publications - Vatsal Kedia
Indian Institute of Technology Bombay
2 shared publications
Latest Publications
- Open Access
Incremental Data Driven Transfer Identification
We introduce a geometric method for online transfer identification of a deterministic linear time-invariant system. At the beginning of the identification process, we assume access to abundant data from a system that is similar, though not identical, to the true system. In the early stages of data collection from the true system, the dataset generated is still not sufficiently informative to enable precise identification. Consequently, multiple candidate models remain consistent with the observations available at that point. Our method picks, at each instant, the model closest to the similar system that is consistent with the current data. As more data are collected, the proposed model gradually moves away from the initial similar system and eventually converges to the true system when the data set grows to be informative. Numerical examples demonstrate the effectiveness of the incremental transfer identification paradigm, where identified models with minimal data are used to solve the pole placement problem.
Open AccessIncremental Data Driven Transfer Identification
We introduce a geometric method for online transfer identification of a deterministic linear time-invariant system. At the beginning of the identification process, we assume access to abundant data from a system that is similar, though not identical, to the true system. In the early stages of data collection from the true system, the dataset generated is still not sufficiently informative to enable precise identification. Consequently, multiple candidate models remain consistent with the observations available at that point. Our method picks, at each instant, the model closest to the similar system that is consistent with the current data. As more data are collected, the proposed model gradually moves away from the initial similar system and eventually converges to the true system when the data set grows to be informative. Numerical examples demonstrate the effectiveness of the incremental transfer identification paradigm, where identified models with minimal data are used to solve the pole placement problem.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Control Systems Letters
- Vol. 10, 1027-1032
Incremental Data Driven Transfer Identification and Mode Detection for Switched Linear Systems
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Control Systems Letters
- Vol. 10, 1027-1032
Incremental Data Driven Transfer Identification and Mode Detection for Switched Linear Systems
- 7 April 2025
- Published by Taylor & Francis in International Journal of Control
- Vol. 98 (12) , 2835-2846
Randomised subspace system identification: complexities and error bounds
- 7 April 2025
- Published by Taylor & Francis in International Journal of Control
- Vol. 98 (12) , 2835-2846
Randomised subspace system identification: complexities and error bounds
- 17 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Control Systems Letters
- Vol. 8, 3416-3421
Fast Data-Driven Predictive Control for LTI Systems: A Randomized Approach
In this letter, the problem of reducing the computational complexity of a recently developed data-driven predictive control scheme is considered. For this purpose, a randomized data compression technique is proposed, which makes the dimension of the decision variable independent of the recorded data size, thereby reducing the complexity of the online optimization problems in data-driven predictive control to that of classical model-based predictive control. The proposed method outperforms other competing complexity reduction schemes in benchmark tests, while guaranteeing similar control performance and stability properties.
- 17 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Control Systems Letters
- Vol. 8, 3416-3421
Fast Data-Driven Predictive Control for LTI Systems: A Randomized Approach
In this letter, the problem of reducing the computational complexity of a recently developed data-driven predictive control scheme is considered. For this purpose, a randomized data compression technique is proposed, which makes the dimension of the decision variable independent of the recorded data size, thereby reducing the complexity of the online optimization problems in data-driven predictive control to that of classical model-based predictive control. The proposed method outperforms other competing complexity reduction schemes in benchmark tests, while guaranteeing similar control performance and stability properties.
- 13 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Random Forests for Time Optimal Feedback Control and Pursuit Evasion Games
In this paper, the suitability of decision tree/random forest classifiers for learning the feedback solutions to time-optimal control problems and pursuit evasion games involving Dubins vehicles, is investigated through several numerical experiments. Recent results [1] have shown that feedback solutions to the game of two cars are bang-bang with the pursuer and evader controls switching between extreme values based on the relative geometry of the players. However no explicit formula for the feedback laws exist. This article demonstrates that this geometry based logic can be learnt by decision trees/forests quite efficiently, resulting in relatively simple implementations of the feedback laws. The laws thus learned matches with analytical solutions for the variants of the game where these are known. For the cases where such closed form solutions are unknown, the decision tree/forest based laws match with optimal trajectories computed according to other available numerical methods.
- 13 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Random Forests for Time Optimal Feedback Control and Pursuit Evasion Games
In this paper, the suitability of decision tree/random forest classifiers for learning the feedback solutions to time-optimal control problems and pursuit evasion games involving Dubins vehicles, is investigated through several numerical experiments. Recent results [1] have shown that feedback solutions to the game of two cars are bang-bang with the pursuer and evader controls switching between extreme values based on the relative geometry of the players. However no explicit formula for the feedback laws exist. This article demonstrates that this geometry based logic can be learnt by decision trees/forests quite efficiently, resulting in relatively simple implementations of the feedback laws. The laws thus learned matches with analytical solutions for the variants of the game where these are known. For the cases where such closed form solutions are unknown, the decision tree/forest based laws match with optimal trajectories computed according to other available numerical methods.
- 3 June 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automatic Control
- Vol. 67 (4) , 1806-1821
A Time-Optimal Feedback Control for a Particular Case of the Game of Two Cars
In this paper, a computationally efficient time-optimal feedback solution to the game of two cars, for the case where the pursuer is faster and more agile than the evader, is presented. The concept of continuous subsets of the reachable set is introduced to characterize the time-optimal pursuit-evasion game under feedback strategies. Using these subsets it is shown that, if initially the pursuer is distant enough from the evader, then the feedback saddle point strategies for both the pursuer and the evader are coincident with one of the common tangents from the minimum radius turning circles of the pursuer to the minimum radius turning circles of the evader. Using geometry, four feasible tangents are identified and the feedback min-max strategy for the pursuer and the max-min strategy for the evader are derived by solving a 2 2 matrix game at each instant. Insignificant computational effort is involved in evaluating the pursuer and evader inputs using the proposed feedback control law and hence it is suitable for real-time implementation.
- 3 June 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automatic Control
- Vol. 67 (4) , 1806-1821
A Time-Optimal Feedback Control for a Particular Case of the Game of Two Cars
In this paper, a computationally efficient time-optimal feedback solution to the game of two cars, for the case where the pursuer is faster and more agile than the evader, is presented. The concept of continuous subsets of the reachable set is introduced to characterize the time-optimal pursuit-evasion game under feedback strategies. Using these subsets it is shown that, if initially the pursuer is distant enough from the evader, then the feedback saddle point strategies for both the pursuer and the evader are coincident with one of the common tangents from the minimum radius turning circles of the pursuer to the minimum radius turning circles of the evader. Using geometry, four feasible tangents are identified and the feedback min-max strategy for the pursuer and the max-min strategy for the evader are derived by solving a 2 2 matrix game at each instant. Insignificant computational effort is involved in evaluating the pursuer and evader inputs using the proposed feedback control law and hence it is suitable for real-time implementation.
- Open Access
A time-optimal feedback control for a particular case of the game of two cars
In this paper, a time-optimal feedback solution to the game of two cars, for the case where the pursuer is faster and more agile than the evader, is presented. The concept of continuous subsets of the reachable set is introduced to characterize the time-optimal pursuit-evasion game under feedback strategies. Using these subsets it is shown that, if initially the pursuer is distant enough from the evader, then the feedback saddle point strategies for both the pursuer and the evader are coincident with one of the common tangents from the minimum radius turning circles of the pursuer to the minimum radius turning circles of the evader. Using geometry, four feasible tangents are identified and the feedback min-max strategy for the pursuer and the max-min strategy for the evader are derived by solving a $2 \times 2$ matrix game at each instant. Insignificant computational effort is involved in evaluating the pursuer and evader inputs using the proposed feedback control law and hence it is suitable for real-time implementation. The proposed law is validated further by comparing the resulting trajectories with those obtained by solving the differential game using numerical techniques.Open AccessA time-optimal feedback control for a particular case of the game of two cars
In this paper, a time-optimal feedback solution to the game of two cars, for the case where the pursuer is faster and more agile than the evader, is presented. The concept of continuous subsets of the reachable set is introduced to characterize the time-optimal pursuit-evasion game under feedback strategies. Using these subsets it is shown that, if initially the pursuer is distant enough from the evader, then the feedback saddle point strategies for both the pursuer and the evader are coincident with one of the common tangents from the minimum radius turning circles of the pursuer to the minimum radius turning circles of the evader. Using geometry, four feasible tangents are identified and the feedback min-max strategy for the pursuer and the max-min strategy for the evader are derived by solving a $2 \times 2$ matrix game at each instant. Insignificant computational effort is involved in evaluating the pursuer and evader inputs using the proposed feedback control law and hence it is suitable for real-time implementation. The proposed law is validated further by comparing the resulting trajectories with those obtained by solving the differential game using numerical techniques. - 21 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automatic Control
- Vol. 66 (10) , 5013-5020
Intermittent Feedback Control With Maximum Average Off-Time
An intermittent feedback-based control policy is proposed for rejecting bounded disturbance signals in a continuous, linear time-invariant (LTI) system with distinct and rational eigenvalues. The proposed policy maximizes the average feedback-off time while retaining the state trajectory inside a prespecified safe region. To achieve this, the feedback link is alternatively turned on and off. During feedback-on intervals, the state trajectory is steered to the origin in min–max time and then feedback is turned off for some maximal precomputed duration. The largest subset of the safe region that is invariant under the proposed policy is characterized. The feedback control and the open-loop control in the proposed policy, which are applied during feedback-on and feedback-off intervals, respectively, are obtained by solving certain min–max and max–min time-optimal control problems. It is shown that the proposed policy achieves the abovementioned objective of maximizing the average feedback-off time. An interesting additional feature of this policy is that the optimal control problems in it need to be solved only once offline and, hence, they do not increase the burden of real-time computation.
- 21 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automatic Control
- Vol. 66 (10) , 5013-5020
Intermittent Feedback Control With Maximum Average Off-Time
An intermittent feedback-based control policy is proposed for rejecting bounded disturbance signals in a continuous, linear time-invariant (LTI) system with distinct and rational eigenvalues. The proposed policy maximizes the average feedback-off time while retaining the state trajectory inside a prespecified safe region. To achieve this, the feedback link is alternatively turned on and off. During feedback-on intervals, the state trajectory is steered to the origin in min–max time and then feedback is turned off for some maximal precomputed duration. The largest subset of the safe region that is invariant under the proposed policy is characterized. The feedback control and the open-loop control in the proposed policy, which are applied during feedback-on and feedback-off intervals, respectively, are obtained by solving certain min–max and max–min time-optimal control problems. It is shown that the proposed policy achieves the abovementioned objective of maximizing the average feedback-off time. An interesting additional feature of this policy is that the optimal control problems in it need to be solved only once offline and, hence, they do not increase the burden of real-time computation.
- Open Access
Estimation of undetected COVID-19 infections in India
Background and Objectives: While the number of detected COVID-19 infections are widely available, an understanding of the extent of undetected COVID-19 cases is urgently needed for an effective tackling of the pandemic and as a guide to lifting the lockdown. The aim of this work is to estimate and predict the true number of COVID-19 (detected and undetected) infections in India for short to medium forecast horizons. In particular, using publicly available COVID-19 infection data up to 28th April 2020, we forecast the true number of infections in India till the end of lockdown (3rd May) and five days beyond (8th May). Methods: The high death rate observed in most COVID-19 hit countries is suspected to be a function of the undetected infections existing in the population. An estimate of the age weighted infection fatality rate (IFR) of the disease of 0.41%, specifically calculated by taking into account the age structure of Indian population, is already available in the literature. In addition, the recorded case fatality rate (CFR= 1%) of Kerala, the first state in India to successfully flatten the curve by consistently reporting single digit new infections from 12-20 April, is used as a second estimate of the IFR. These estimates are used to formulate a relationship between deaths recorded and the true number of infections and recoveries. The estimated undetected and detected cases time series based on these two IFR estimates are then used to fit a discrete time multivariate infection model to predict the total infections at the end of the formal lockdown period. Results: Over three consecutive fortnight periods during the lockdown, it was noted that the rise in detected infections has decreased by 8.2 times. For an IFR of 0.41%, the rise in undetected infections decreased 2.5 times, while for the higher IFR value of 1%, undetected cases decreased by 2.4 times. The predicted number of total infections in India on 3rd May for both IFRs varied from 2.8 - 6.8 lakhs. Interpretation and Conclusions: The behaviour of the undetected cases over time effectively illustrates the effects of lockdown and increased testing. From our estimates, it is found that the lockdown has brought down the undetected to detected cases ratio, and has consequently dampened the increase in the number of total cases. However, even though the rate of rise in total infections has fallen, the lifting of the lockdown should be done keeping in mind that 2.3 to 6.4 lakhs undetected cases will already exist in the population by 3rd May.
Open AccessEstimation of undetected COVID-19 infections in India
Background and Objectives: While the number of detected COVID-19 infections are widely available, an understanding of the extent of undetected COVID-19 cases is urgently needed for an effective tackling of the pandemic and as a guide to lifting the lockdown. The aim of this work is to estimate and predict the true number of COVID-19 (detected and undetected) infections in India for short to medium forecast horizons. In particular, using publicly available COVID-19 infection data up to 28th April 2020, we forecast the true number of infections in India till the end of lockdown (3rd May) and five days beyond (8th May). Methods: The high death rate observed in most COVID-19 hit countries is suspected to be a function of the undetected infections existing in the population. An estimate of the age weighted infection fatality rate (IFR) of the disease of 0.41%, specifically calculated by taking into account the age structure of Indian population, is already available in the literature. In addition, the recorded case fatality rate (CFR= 1%) of Kerala, the first state in India to successfully flatten the curve by consistently reporting single digit new infections from 12-20 April, is used as a second estimate of the IFR. These estimates are used to formulate a relationship between deaths recorded and the true number of infections and recoveries. The estimated undetected and detected cases time series based on these two IFR estimates are then used to fit a discrete time multivariate infection model to predict the total infections at the end of the formal lockdown period. Results: Over three consecutive fortnight periods during the lockdown, it was noted that the rise in detected infections has decreased by 8.2 times. For an IFR of 0.41%, the rise in undetected infections decreased 2.5 times, while for the higher IFR value of 1%, undetected cases decreased by 2.4 times. The predicted number of total infections in India on 3rd May for both IFRs varied from 2.8 - 6.8 lakhs. Interpretation and Conclusions: The behaviour of the undetected cases over time effectively illustrates the effects of lockdown and increased testing. From our estimates, it is found that the lockdown has brought down the undetected to detected cases ratio, and has consequently dampened the increase in the number of total cases. However, even though the rate of rise in total infections has fallen, the lifting of the lockdown should be done keeping in mind that 2.3 to 6.4 lakhs undetected cases will already exist in the population by 3rd May.
- 20 December 2019
- Published by Taylor & Francis in International Journal of Control
- Vol. 94 (8) , 2282-2295
Time optimal consensus tracking with multiple leaders
This paper considers a multi-agent system with multiple leader agents. All the agents have bounded inputs. Any two agents can communicate only if they are within a fixed distance of each other. Leaders are special agents which generate trajectories autonomously based upon certain objectives. The follower agents must partition into groups such that agents in each group track exactly one leader. Given a particular partition and the resulting information exchange topology, every follower in each group must converge onto their leader's trajectory in min-max time. A special directed spanning tree is selected in each group. Each follower agent in this tree time-optimally pursues its tail agent using a feedback policy. It is shown that these local feedback laws achieve global min-max time consensus tracking for any general information graph. Further, a decentralized asynchronous algorithm is proposed to optimally partition the agents so that consensus tracking for all the groups is accomplished in the shortest possible time.
- 20 December 2019
- Published by Taylor & Francis in International Journal of Control
- Vol. 94 (8) , 2282-2295
Time optimal consensus tracking with multiple leaders
This paper considers a multi-agent system with multiple leader agents. All the agents have bounded inputs. Any two agents can communicate only if they are within a fixed distance of each other. Leaders are special agents which generate trajectories autonomously based upon certain objectives. The follower agents must partition into groups such that agents in each group track exactly one leader. Given a particular partition and the resulting information exchange topology, every follower in each group must converge onto their leader's trajectory in min-max time. A special directed spanning tree is selected in each group. Each follower agent in this tree time-optimally pursues its tail agent using a feedback policy. It is shown that these local feedback laws achieve global min-max time consensus tracking for any general information graph. Further, a decentralized asynchronous algorithm is proposed to optimally partition the agents so that consensus tracking for all the groups is accomplished in the shortest possible time.