Yasra Chandio
I have been involved in several projects particularly in the field of networked embedded systems, mixed reality, and security. Some of these projects include:
Ongoing Projects
Extended Reality (XR)/Mixed Reality (MR) Security: In this work, we exhaustively explore the threat landscape of simple, fast, and stealthy attacks in XR/MR systems. This attack surface is resilient to state-of-the-art sensor fusion techniques by concurrently manipulating multiple sensing streams across multiple dimensions. A major focus of this work is to establish stealthiness in the presence of sophisticated security mechanisms with an added constraint that XR/MR systems allow minimal time durations for covert operations. The efficacy of proposed attacks is evaluated through a preliminary study (Poster Abstract at ACM SenSys'20) on inertial and visual data streams. (Note: Full work is under review at SenSys 2022.)
Quantifying Presence in Mixed Reality: Presence is the feeling of being physically and spatially located in an environment. Presence degrades when there are inconsistencies in the virtual environment. Our work aims to use response delay to quantify the user presence. To do that, we have established from the literature that presence impacts response delay. Once we establish this correlation, we quantify presence using the response delay data.
Domain Adaptive Tracking in Mixed Reality: Mixed reality applications require tracking systems with high accuracy, low latency, and low jitter. These systems must recover the full six degrees of freedom (6 DOF) of the user’s head pose— the position and orientation of the user’s head in the scene coordinate system. While several available solutions exist for stationary and indoor tracking, a limited volume of solutions exists for mobile and outdoor MR systems. This work aims to “robustify” the sensor fusion algorithms by making the associated deep learning models more explainable and exploring solutions such as domain randomization in simulation to robustify different feature-encoders. The idea is that the proposed approach should be robust across diverse contexts, noise sources, and sensor types while using the context of a scene to improve overall performance.
Past Projects
Hawadaar: An IoT-based retrofitted HVAC system for personalized comfort that provides centralized control abstraction to air-condition efficiently traditional buildings lack HVAC and use distributed legacy devices such as heater and AC units. Hawadaar can save up to 6% of electricity per capita in residential and commercial sectors — resulting in a substantial countrywide impact in Pakistan. (ACM BuildSys Paper)
WET Simulator: A generic TOSSIM-based simulation framework to model energy harvesting (EH) and wireless energy transfer (WET) that enables rapid development of harvesting and transfer-aware applications, protocols, and system software in wireless sensor networks (WSN). The framework easily integrates new models providing a potential platform for the WSN community to evaluate sophisticated models of various EH and WET technologies. (EWSN Paper)
Energy Stack: A layered software model for energy management in transiently-powered wireless embedded networks to create a global energy view that facilitates network-wide optimization and management of energy. We delineate the responsibilities of different layers in the energy stack, define interfacing APIs between layers, identify corresponding challenges, and provide a first implementation of the energy stack. (TOSN paper)
iCompute: A partial check-pointing solution for IoT system software; enabling data networking under intermittent energy supply, with coordinated check-pointing for distributed applications in IoT. The key contribution of the project is incremental checkpointing techniques enabling transiently powered devices to retain computational state across multiple activation cycles.