Urbashi Mitra - Selected Publications#

J. Li, M. F. Da Costa and U. Mitra, "Joint Localization and Orientation Estimation in Millimeter-Wave MIMO OFDM Systems via Atomic Norm Minimization," in IEEE Transactions on Signal Processing, vol. 70, pp. 4252-4264, 2022.

This work provided a novel tensor signal representation that enabled the use of powerful atomic norm methods for localization, a service that is critical to the Internet-of-Things (IoT), navigation, and safety applications. Conditions for high performance recovery of the location are provided and the proposed method is shown to be near-optimal. The paper has spawned new work in ensuring the privacy of location to eavesdroppers.

M. Choraria, A. Chattopadhyay, U. Mitra and E. G. Ström, "Design of False Data Injection Attack on Distributed Process Estimation," in IEEE Transactions on Information Forensics and Security, vol. 17, pp. 670-683, 2022

Understanding how to be resilient against network attacks is critical for modern applications employing wireless communications such as V2V or IoT networks. This paper develops a novel, low complexity on-linear learning scheme to optimize the attack. Modern methods of stochastic approximation were exploited. The new attacks deviate from classical structures and, counter-intuitively, are not random. Understanding attack methods enables effective countermeasures.

A. Chattopadhyay and U. Mitra, "Security Against False Data-Injection Attack in Cyber-Physical Systems," in IEEE Transactions on Control of Network Systems, vol. 7, no. 2, pp. 1015-1027, June 2020

A new strategy for determining compromised agents in a multi-agent system is proposed that is both low complexity and of low communication cost. Agents only need to communicate with neighbors to determine the presence of an attack. Theoretical properties using a challenging two-time scale analysis are proven – in fact, the approach is provably good.

G. Aminian, H. Arjmandi, A. Gohari, M. Nasiri-Kenari and U. Mitra, "Capacity of Diffusion-Based Molecular Communication Networks Over LTI-Poisson Channels," in IEEE Transactions on Molecular, Biological and Multi-Scale Communications, vol. 1, no. 2, pp. 188-201, June 2015.

Molecular communications is an enabling technology for in-body health sensing and drug delivery. Determining the maximum amount of information that can be transmitted for such systems has proven to be quite challenging due the complexity of these channels. This work exploits a novel channel description and proposes a new bound on capacity that is straightforward to compute and effective in the small signal regime.

M. Levorato, U. Mitra and M. Zorzi, "Cognitive Interference Management in Retransmission-Based Wireless Networks," in IEEE Transactions on Information Theory, vol. 58, no. 5, pp. 3023-3046, May 2012.

Cognitive radio systems enable agile, ad hoc wireless users to take advantage of underutilized wireless resources. Prior to this work, the bulk of the cognitive radio research focused on ensuring that the agile, cognitive users completely avoided existing, legacy users. Furthermore, legacy devices were assumed to be quite simple in contrast to the devices being deployed. This paper showed that by allowing for a little bit of interference to legacy devices, cognitive devices could strongly improve their performance even when the legacy devices change their communication strategies in the presence of the cognitive users.

D. -S. Zois, M. Levorato and U. Mitra, "Energy-Efficient, Heterogeneous Sensor Selection for Physical Activity Detection in Wireless Body Area Networks," in IEEE Transactions on Signal Processing, vol. 61, no. 7, pp. 1581-1594, April 2013.

Informed by a real-world deployment of a wireless body area sensing network for health telemetry, a novel sensor allocation scheme was developed. In contrast to prior (idealized) work, heterogeneous sensors wherein the fusion center (mobile phone) was the energy bottleneck was considered. An optimal, Kalman-like filter was derived; more importantly, by exploiting properties of the optimal solution, a low complexity sensor selection strategy was developed and validated on real-world data. The low complexity strategies perform near optimally.

Kuan-Wen Huang, M. Bică, U. Mitra and V. Koivunen, "Radar waveform design in spectrum sharing environment: Coexistence and cognition," 2015 IEEE Radar Conference (RadarCon), Arlington, VA, USA, 2015, pp. 1698-1703.

This paper was an early work in the now-crowded field of sensing and communication in mixed radar/communication environments. While prior work had focused on either the communication goal or the radar goal, this paper adopts a joint consideration. To this end, optimal radar waveforms are designed via optimizing information theoretic metrics such that a minimum capacity is maintained for the communication system. A dual problem is considered for optimization target detection wherein the solution is a spectrally efficient OFDM radar signal which performs better than classical OFDM (widely used in WiFi and WLANs).

W. Zhang, M. Stojanovic and U. Mitra, "Analysis of a Linear Multihop Underwater Acoustic Network," in IEEE Journal of Oceanic Engineering, vol. 35, no. 4, pp. 961-970, Oct. 2010.

Underwater acoustic (UA) communications are needed for environmental monitoring, harbor security and aquaculture. Given the significant signal attenuation in UA channels, using relays to enable long distance communication is necessary. This paper exploits the unique characteristics of the UA channel to design a novel time-space-power allocation scheme to maximize the amount of information which can be transmitted. The unique nature of half-duplex radios and communications with error-correcting coding are considered for an information theoretic analysis.

R. Nowak, U. Mitra and R. Willett, "Estimating inhomogeneous fields using wireless sensor networks," in IEEE Journal on Selected Areas in Communications, vol. 22, no. 6, pp. 999-1006, Aug. 2004.

While image-processing types of approaches had been employed for boundary detection in sensor networks, they were ad hoc in nature and challenging to analyze. This work exploited hierarchical representations of images that enabled the direct connection between the performance of boundary detection and communication costs between sensors. In particular, a theoretical analysis of the tradeoffs between performance and communication as a function of the density of the nodes was provided. A practical algorithm was designed that mapped well to sensor network architectures and was proven to achieve the best-case performance/energy tradeoff.

G. Hollinger, U. Mitra, and G. Sukhatme, 'Active classification: Theory and application to underwater inspection,' In Robotics Research: The 15th International Symposium, pp. 95-110, 2017.

This work highlights Mitra’s interdisciplinary work at the intersection of communications, control and sensing. An underwater autonomous must classify an object based on multiple views. Using a key adaption of Bayesian active learning a new active classification method is proposed which is probabilistic in nature exploiting information theoretic cost functions. The numerical results (from real and synthetic data) show that the new algorithm for actively planning for informative views can reduce the number of necessary views by up to 80 % when compared to prior methods.

The above selected papers are very well cited. See the Google Scholar page of Dr. Urbashi Mitra for detailed information on citations and publications.