Research

General Area: Data Analytics, Autonomous Learning Systems for Dynamically Evolving Environments ,, Computational Intelligence, Big Data

Sub-areas: Anomaly Detection, Autonomous Video-Analytics, Mining Data Streams in Real time, Biomedical applications, Robotic and Autonomous Vehicles, CyberSecurity,Evolving User Behaviour Modelling, Human Activity Recognition, Ambient Assisted Living, Security and Surveillance, Defense, etc.

Theoretical contributions

TEDA

evolving classifiers family eClass, AutoClass

Evolving Systems - a new concept, methods, algorithms, tools, 2000-now; see also Plamen Angelov (2008) Evolving fuzzy systems. Scholarpedia, 3(2):6274 as well as the Springer journal EVOS

Recursive Density Estimation (RDE) - 2003-now - a recursive version of the well known KDE (kernel density estimation) using Cauchy kernel (instead of Gaussian), see also P. Angelov, Anomalous System State Identification, patent application, GB1208542.9, priority date 15 May 2012.

Rule-based models with non-parametric antecedents (type ANYA) - a step beyond Takagi-Sugeno models (2010-2011) with Prof. Ronald R Yager; ANYA classifiers, predictors, controllers (there were two well known types of fuzzy systems: Mamdani and Takagi-Sugeno; AnYa claims to be the third alternative simpler yet more efficient and easier to design type), see also P. Angelov, R.Yager, A New Type of Simplified Fuzzy Rule-based Systems, International Journal of General Systems, vol.41(2), 163-185, Jan. 2012.

Optimal Online Fisher self-learning classifier (2011) with Denis Kolev

Genetic Algorithm- assisted design of Fuzzy Systems (GFS), 1996-2000

Automatic Configuration Generator, 1999-2003
GA/EA-based novel approach to automate the whole process of design of component-based systems (in particular HVAC) starting from the design brief, all available data about the environment, lnfeasible inks of components, their physical parameters, and the constraints and requirements

New operators for Genetic Algorithms - centre-of-gravity-based crossover, 1998-2000

Fuzzy Mathematical Programming and Generalzed Fuzzy Optimization - analytical solutions, 1992-1995

Fuzzy Optimal Control - a new approach to optimal control under uncertainties, 1990-1993, 2006-(application to adaptive routeing)

Applied research

Demos

  • Assisted Carriage - Intelligent Leader - follower algorithms for ground platforms (1min 14sec demo)' MoD project RT/COM/7/047 (student's poster)
  • Assisted Carriage - indoor experiment (47 sec. video)
  • demo1 v1, demo1_v2 (4 min demos)
  • obstacle avoidance module_v1,- obstacle avoidance module_v2 (15 sec demos)
  • Robot follower on youtube (4 min demo)
  • Identifying motion status of the leader (23 sec demo)
  • Detecting the 'leader' by superposition of scans and clustering (11sec video demo)
  • evolving Clustering for Remote monitoring and pattern recognition in the state of a patient (or a soldier) - Person Monitoring using Wireless Sensor Network (3 min video)
  • Landmarks automatic recognition by an autonomous vehicle - S(elf)-L(ocalisation) A(nd) M(apping)
  • 'Follow the leader' experiment - 'intelligent mule' - 2006 version
  • Detecting (using R(ecursive) D(ensity) E(stimator)) and tracking (using evolving Takagi-Sugeno, eTS type systems) targets in noisy (rainy) environment video
  • Detecting (using RDE), identifying (using evolving Clustering) and tracking (using eTS) multiple targets - original video; detection and tracking