Browsing by Author "Ravindu, HM"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    item: Conference-Full-text
    Cloud movement analysis using Lucas Kanade method
    (Institute of Electrical and Electronics Engineers, Inc., 2021-09) Weligamage, SU; Ravindu, HM; Bandara, T; Ranathunga, L; Abeykoon, AMHS; Velmanickam, L
    Weather forecasting is an important task as it helps people to manage their work.Cloud movement is a major factor for the variation of the weather. Along with the other factors, cloud movement scan beused to predict the weather changes. This paper propose a method to analyze the cloud movements and calculate the motion vectors using Lucas Kanade method. Cloud velocity vector scan beused with cloud types, wind and thermal data to predict the cloud patterns which helps to forecast weather changes. Proposed method has implemented on a software plat form and experiments has been carried out to validate the system.
  • Thumbnail Image
    item: Conference-Full-text
    User Profiling Based Proactive Interaction Manager for Adaptive Human-robot Interaction
    (IEEE, 2023-12-09) Rajendran, H; Ravindu, HM; Bandara, T; Jayasekara, AGBP; Chandima, DP; Abeysooriya, R; Adikariwattage, V; Hemachandra, K
    Proactive interaction is a highly researched topic in the robotics arena. By using its proactive ability, a robot can decide intelligently in specific situations, taking particular attributes into account, to offer a certain service to the user. We propose a proactive interaction manager to suggest suitable reading suggestions to the user. User profiling is used as a tool to identify the nature and preferences of each user. A user’s profile is based on multiple attributes that determine whether a user chooses a particular genre for reading. This is developed as a strategy to identify the preferences of users. Since the user’s preferences may change over time, an algorithm to adaptively provide reading suggestions to the user considering their changes in preferences is also proposed. We conducted two experiments to evaluate the performance of the proposed proactive interaction manager. The results of the experiments show that the proposed approach is satisfactory for a domestic service robot based application.