Research

The coming decades may see the large scale deployment of networked cyber-physical systems to address global needs in areas such as energy, water, healthcare, and transportation. However, as recent events have shown, such systems are vulnerable to cyber attacks. Being safety critical, their disruption or misbehavior can cause economic losses or injuries and loss of life. It is therefore important to secure such networked cyber-physical systems against attacks. In the absence of credible security guarantees, there will be resistance to the proliferation of cyber-physical systems, which are much needed to meet global needs in critical infrastructures and services.

This line of work addresses the problem of secure control of networked cyber-physical systems. This problem is different from the problem of securing the communication network, since cyber-physical systems at their very essence need sensors and actuators that interface with the physical plant, and malicious agents may tamper with sensors or actuators, as recent attacks have shown.

We consider physical plants that are being controlled by multiple actuators and sensors communicating over a network, where some sensors and actuators could be "malicious," as indicated in the accompanying figure. A malicious sensor may not report the measurements that it observes, and a malicious actuator may not apply inputs as specified by the control policy. Strategic collusion of such malicious sensors and actuators could destabilize the closed-loop system or deteriorate its performance. 

In one line of work, we address a general technique by which the honest actuators and sensors in the system can detect the actions of malicious nodes, and disable closed-loop control based on their information. This technique, called "watermarking," employs the technique of actuators injecting private excitation into the system which will reveal malicious tampering with signals. We show how such an active defense can be used to secure networked systems of sensors and actuators.

The notions of controllable and unobservable subspaces introduced by Kalman have played a central role in the study of modern control theory. These notions, however, were introduced in an era when all sensors and actuators in a system could be trusted, and all disturbances affecting the system are stochastic. In another line of work, we introduce the notions of securable and unsecurable subspaces of a linear system, and show that they have important operational meanings in the context of secure control. These subspaces can be regarded as the analogs of controllable and unobservable subspaces in an era where there is intense interest in cybersecurity of control systems. 

Papers on CPS Security:

[1] B. Satchidanandan, and P. R. Kumar, "Dynamic Watermarking: Active Defense of Networked Cyber-Physical Systems," in Proceedings of the IEEE , vol. 105, no. 2, pp.219-240, February 2017.

[2] B. Satchidanandan, and P. R. Kumar, "Secure Control of Networked Cyber-Physical Systems," 2016 IEEE 55th Conference on Decision and Control (CDC), Las Vegas, NV, USA, 2016, pp. 283-289. 

[3] Woo-Hyun Ko, B. Satchidanandan, and P. R. Kumar, "Theory and Implementation of Dynamic Watermarking for Cybersecurity of Advanced Transportation Systems," to appear in the Proceedings of the 2016 International Workshop on Cyber-Physical Systems Security (CPS Sec.), Philadelphia, USA, October 2016, pp. 235-239. 

[4] B. Satchidanandan, and P. R. Kumar, "On Minimal Tests of Sensor Veracity for Dynamic Watermarking-Based Defense of Cyber-Physical Systems," to appear in the 9th International Conference on Communication Systems and Networks (COMSNETS), Bangalore, India, January 2017. (Best Student Paper Award) 

[5] B. Satchidanandan and P. R. Kumar, "Defending Cyber-Physical Systems from Sensor Attacks," Lecture Notes in Computer Science, Vol. 10340, pp. 150-176, Springer International Publishing, September 2017.

[6] B. Satchidanandan and P. R. Kumar, "Control Systems Under Attack: The Securable and Unsecurable Subspaces of a Linear Stochastic System," Emerging Applications of Control and Systems Theory, Springer International Publishing, 2018, pp. 217-228.

[7] B. Satchidanandan and P. R. Kumar, "The securable subspace of a linear stochastic system with malicious sensors and actuators," 2017 55th Annual Allerton Conference on Communication, Control, and Computing (Allerton), Monticello, IL, 2017, pp. 911-917.

[8] T. Huang, B. Satchidanandan, P. R. Kumar and L. Xie, "An Online Detection Framework for Cyber Attacks on Automatic Generation Control," in IEEE Transactions on Power Systems, vol. 33, no. 6, pp. 6816-6827, Nov. 2018.

[9] B. Satchidanandan and P. R. Kumar, "On the Operational Significance of the Securable Subspace for Partially Observed Linear Stochastic Systems," 2018 IEEE Conference on Decision and Control (CDC), Miami Beach, FL, 2018, pp. 2068-2073.

[10] B. Satchidanandan and P. R. Kumar, "On the Design of Security-Guaranteeing Dynamic Watermarks," in IEEE Control Systems Letters, vol. 4, no. 2, pp. 307-312, April 2020.

[11] B. Satchidanandan and P. R. Kumar, "On the Watermark-Securable Subspace of a Linear Stochastic System," in Proceedings of the 2019 Indian Control Conference (ICC), Hyderabad, India. (Finalist, Best Student Paper Award)

[12] Woo-Hyun Ko, B. Satchidanandan, and P. R. Kumar, "Dynamic Watermarking-Based Defense of Transportation Cyber-Physical Systems," in ACM Transactions on Cyberphysical Systems. 

[13] B. Satchidanandan and P. R. Kumar, "The Watermark-Securable Subspace of a Linear Stochastic System Containing a Single Malicious Actuator," in Proceedings of the 2020 International Conference on Communication Systems and Networks, January 2020.

Cyber-Physical-Social Systems (CPSS) refer to Cyber-Physical Systems with human beings in the loop. The behavior of a CPSS is governed not only by physical laws, but also by the behaviors of human beings that interact with the system. The power system and the transportation system are two prime examples. A holistic paradigm of controlling such systems should involve controlling not only the physical signals that actuate the system, but also the behaviors of human beings that interact with the system. The high-level goal of this line of research is to extend mechanism design theory --- a subfield of microeconomics/game theory which has a long history of addressing problems related to influencing human behavior --- to bear upon problems arising in the context of controlling cyber-physical-social systems.

In our prior work, we have addressed certain special cases of the general problem described above. Specifically, we have introduced in [1] the environment of a Two-Stage Repeated Stochastic Game using which many emerging problems in power systems can be readily modeled, and have devised an incentive mechanism to drive human behavior toward a desired equilibrium in such environments. In other works, we have applied the theory developed in [1] to solve two important problems that arise in the context of controlling next-generation power grids: EV-Grid Integration and Demand Response. 

Demand Response

The first problem pertains to Demand Response. One of the characteristic features of the power grid is that it has very little storage, and so any power that is produced must be immediately consumed. This imposes the constraint that the demand  and supply must be equal at every instant, not just on average. The way such balance is maintained today is by adjusting the generation to follow the random demand. This is possible as power generation today is controllable. However, in a futuristic power grid with high share of renewables in the generation portfolio, the supply will no longer be completely controllable. A popular paradigm for maintaining demand-supply balance in such a scenario is Demand Response, which refers to making the demand follow the random supply. This is achieved by offering financial incentives to users to curtail power consumption when there is supply shortage and increase consumption when the supply is in excess. Simple though the high-level idea may appear, a careful examination reveals certain difficulties which stem from informational asymmetries between Demand Response providers and the System Operator. Our work [2] applies the mechanism developed in [1] to address these issues. 

EV-Grid Integration

The second problem pertains to Electric Vehicle-Power Grid integration. As mentioned before, a key obstacle to increasing renewable energy penetration in the power grid is the lack of utility-scale storage capacity. Transportation electrification has the potential to overcome this obstacle since Electric Vehicles (EVs) that are not in transit can provide battery storage as a service to the grid. This is referred to as EV-Power grid integration, and could potentially be a key milestone in the pathway to decarbonize the electricity and the transportation sectors. Successful integration of EV storage into the grid requires the accompaniment of a host of reforms in grid and electricity market operations, certain important ones of which we have addressed in [3]-[5]. 

The first reform pertains to redesigning the day-ahead and real-time electricity markets to allow for EVs (or any other energy storage provider for that matter) to participate alongside generators and loads. In [3] and [4], we have developed markets for EVs to sell their battery storage service. 

The second reform pertains to day-ahead market operations. In order to optimally utilize EV storage, the system operator must account for EV storage availability while solving the economic dispatch problem. One of the quintessential features of EV storage that distinguishes it from other forms of energy storage is that EV storage is unreliable: An EV could connect and disconnect from the grid at random times and so the availability of an EV’s battery to the grid is intermittent and random. Consequently, the system operator cannot bank on the energy that it stores in an EV to be available for discharge at a future time. Hence, the economic dispatch problem must account not simply for the availability of storage, but also for the uncertainty in the availability. This results in an economic dispatch problem that is qualitatively different from all the economic dispatch problems that have been addressed in the literature thus far. In [5], we have modeled the unique aspects EV storage and have introduced the EV Storage-Integrated Economic Dispatch Problem, and have presented an optimal approach to solving it.

Papers on CPSS Control:

[1] B. Satchidanandan and Munther A. Dahleh, "Incentive Compatibility in Two-Stage Repeated Stochastic Games," in IEEE Transactions on Control of Networked Systems, To Appear.

[2] B. Satchidanandan, M. Roozbehani, and Munther A. Dahleh, "A Two-Stage Mechanism for Demand Response Markets," in IEEE Control Systems Letters, vol. 7, pp. 49-54, 2023. 

[3] B. Satchidanandan and Munther A. Dahleh, "A Mechanism for Selling Battery Storage Service in Day-Ahead Electricity Markets,'' in Proceedings of the 2021 American Control Conference (ACC), 2021, pp. 2895-2900.

[4] B. Satchidanandan, and Munther A. Dahleh, "An Efficient and Incentive Compatible Mechanism for Energy Storage Markets," in IEEE Transactions on Smart Grid, vol. 13, no. 3, pp. 2245-2258, May 2022.

[5] B. Satchidanandan and Munther A. Dahleh, "Economic Dispatch for EV Energy Storage-Integrated Power Systems,'' in Proceedings of the 14th International Conference on COMmunication Systems & NETworkS (COMSNETS), 2022, pp. 707-715.

Millimeter wave communications constitute an important component of 5G cellular networks. The high bandwidth available at these frequencies aid in overcoming the “spectrum crunch” at lower frequencies. However, certain characteristics of mm-wave radio waves, such as their high blockage, path loss, and oxygen absorption, make the design of mm-wave communication systems radically different from their counterparts operating at sub-6 GHz. Perhaps the most important such difference stems from the fact that highly directional antennae are required for transmission and reception to compensate for the high propagation loss. The high directionality makes nodes “deaf” to neighbors that may be occupying the channel, rendering MAC schemes based on carrier-sense infeasible. In fact, such MAC schemes may even be unnecessary since high directionalities allow for high spatial reuse, as illustrated in the accompanying figure. Therefore, the focus of MAC protocols in this regime must not be on curbing concurrent transmissions, but on coordinating the transmitter and receiver to point their antenna boresights towards each other. This line of work deals with developing such MAC protocols for both ad hoc as well as infrastructure-based indoor wireless networks.

Papers on mm-wave MAC:

[1] B. Satchidanandan, S. Yau, P. R. Kumar, A. Aziz, A. Ekbal and N. Kundargi, "TrackMAC: An IEEE 802.11ad-compatible beam tracking-based MAC protocol for 5G millimeter-wave local area networks," 2018 10th International Conference on Communication Systems & Networks (COMSNETS), Bengaluru, 2018, pp. 185-182. 

[2] B. Satchidanandan, S. Yau, S. S. Ganji, P. R. Kumar, A. Aziz, A. Ekbal and N. Kundargi, "A Directional Medium Access Control Protocol for 5G Millimeter-Wave Local Area Networks," Lecture Notes in Computer Science, Vol. 11227, pp. 150-171, Springer International Publishing, December 2018.  

[3] B. Satchidanandan, V. S. S. Ganji and P. R. Kumar, "Iris: A Directional MAC Protocol With Applications to Millimeter-Wave Mobile Ad-Hoc Networks," 2019 11th International Conference on Communication Systems & Networks (COMSNETS), Bengaluru, India, 2019, pp. 291-297. 

This line of work deals with low-complexity, near-optimal equalization of vector sequences experiencing Inter-Symbol Interference (ISI) and corrupted by Additive White Gaussian Noise (AWGN). Each component of the vector in the vector sequence assumes value from a finite alphabet of possibly "high" cardinality (eg., 16-QAM or 64-QAM). Optimal equalization can be performed at a complexity linear in the sequence length using the well-known Vector Viterbi Algorithm (VVA), but its complexity is superlinear (and for all practical purposes, prohibitive) in the alphabet size, and exponential in the dimension of the vectors. Our research develops approaches for systematic reduction of complexity. 

The mathematical model that we employ is rich enough to model a variety of problems encountered in communications and signal processing, and consequently, the algorithms developed can be applied to a wide class of problems encountered in these fields. These include equalization of (massive) MIMO-ISI channels, channel equalization for systems employing higher-order modulation, multiuser detection in LTE Uplink, co-channel interference suppression in GSM downlink, and receiver design for systems employing faster-than-Nyquist signaling.

Papers on Sequence Detection:

[1] B. Satchidanandan, K. Kuchi, and R. D. Koilpillai, "Generalized Reduced-State Vector Sequence Detection," in IEEE Communications Letters, vol. 18, no. 10, pp. 1691-1694, October 2014.

[2] B. Satchidanandan, K. Kuchi, and R. D. Koilpillai, "Reduced-state soft-output equalization for MIMO-ISI systems employing HOM," Twentieth National Conference on Communications (NCC), Kanpur, 2014, pp. 1-6