Kaijian Liu
Stevens Institute of Technology, NJ·Engineering·Early-career
Disasters disrupt both people and the systems they depend on, such as roads, power grids, water networks and bridges. People change their behavior when infrastructure fails, and infrastructure is in turn affected by what people do.
- Fields
Engineering
Computer Science
- Topics
cyber-physical systems
civil infrastructure
disaster resilience
- Grants
- 2 known grants, through Apr 2031
What they work on
- Develop a learning framework that estimates the conditions of people and infrastructure, across space and time, from sparse and uncertain observations, as a basis for modeling disaster disruptions.
- Build an adaptive learning framework that captures delayed, distributed interactions between human behavior and infrastructure systems such as transportation, power and water networks.
- Create a hybrid simulation and a replicator-inspired multi-agent reinforcement learning method that combine infrastructure changes with uncertain human behavior. The goal is to find stable strategies for people and infrastructure to adapt together.
- Develop deep learning methods informed by engineering knowledge to predict the onset and spread of multiple types of damage in civil infrastructure such as bridges.
- Design reconfigurable sensing based on signal difference maximization theory, plus edge cloud computing for extracting damage information. Test them on bridges using simulations and physical prototypes.
More about the research
Both projects borrow ideas from biology. Species that interact can adapt together, or help each other, to cope with change. The lab applies this idea to cyber-physical systems, which link computing with physical systems, to make infrastructure more resilient. One project models how people and infrastructure adapt to each other during disruptions. The other links damage prediction software with sensors that can be repositioned to monitor bridges more efficiently.
Why it matters
Floods, storms and other disasters disrupt transportation, power and water systems that communities rely on. Models in which people and infrastructure adapt together could help reduce losses and speed recovery after disasters. Reconfigurable sensors guided by damage prediction could let bridge owners monitor several types of damage with fewer sensors. This could lower costs while making damage assessments faster and more accurate.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Deng Cao
Central State University, OH·Engineering
Design a coordinated strategy for connected self-driving vehicles to decide when to send perception tasks to roadside edge computers, keeping the decisions manageable as vehicle numbers grow and robust to uncertain edge behavior.
Vijay Subramanian
University of Michigan, MI·Engineering
Develop robust reinforcement learning algorithms that use model-class knowledge, future predictions and off-policy methods to cope with mismatches and limited data from simulators or offline records.
Aleksandar Stankovic
Santa Clara University, CA·Engineering
Develop techniques combining manifold learning, which finds simple structure in complex data, with information-geometry model reduction to build physically interpretable models that still make quantitative predictions.
The school
Stevens Institute of Technology: 4 labs on PI Pages, $71M a year on research, 72 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.