Swarm intelligence and robotics
Multi-Agent Systems
Temporal networks
Real-time embedded systems
Machine learning and artificial intelligence
Bio-inspired Optimization Techniques
The emergence of collective behavior in biological networks, such as pods of orcas or social networks, is triggered by conditions such as conflict, ambiguous policies, or change in the normative order, all of which is only possible through a sharing and transfer of information. Understanding how information propagates in such networks becomes necessary to build new policies and better technologies. This project targets this issue by modelling the information propagation with minimum access to the system and environment details.
The model is able to properly identify and isolate the originator of specific events and behaviors ; as is the case with giant and snowball effects in rumour propagation research. It detects easily-influenced individuals which are themselves apt to influence other individuals; Hence identifying the major players in information propagation.
The use of simulators and real swarms of robots (Kilobots) to emulate the communication between biological agents validates the proposed model and its findings; more specifically that limited-knowledge incorporated by the model is enough for certain behavior understanding.
In swarm robotics, maintaining the communication among the swarm is very crucial to the overall behavior of the swarm. Insuring convergence towards a certain consensus is one of these desired behaviors. Our aim here is to analyse the timing characteristics of such scenarios by studying the propagation of information in a probabilistic manner. This will open doors to better decision making from the swarm as well as system deployment.
In software simulation development, subject matter experts (SMEs) are hardly expected to translate their expertise in different domains into an imperative programming language, thus leading to the use of DSLs to transfer their knowledge. In the process, information is lost in translation and once at the integration phase, specialist are faced with the ever growing difficult task of putting everything together.
Our goal in this project is to define a methodology to hide software complexity from SMEs and extract the maximum performance from the hardware by taking into account inputs from SMEs and hardware and software experts to automate the software schedulimg, mapping and optimization.
This methodology will be applied in particular to the aerospace industry, and precisely to Full Mission Simulators (FMSs), which are applications that are developed and maintained by multiple SMEs.
I previously worked on implementing an autonomous self-parking car based on SOPC. We built our system on an RC-toy car that detects a parking space, judges wether the space is adequate for parking and then proceeds to park in either parking positions; parallel, or perpendicular. It could also autonomously leave a parking space from a parked position. The system was built on the DE2 Altera FPGA board and used the soft-core technology NIOS-II.