Employing Speeded Scaled Conjugate Gradient Algorithm for Multiple Contiguous Feature Vector Frames: An Approach for Traffic Density State Estimation
Prashant Borkar, Milindkumar V. Sarode, Latesh Malik · Procedia Computer Science · 2016
Modality of employing Neuro-Fuzzy Classifier (ANFC) with Speeded Scaled Conjugate Gradient (SSCG) algorithm for vehicular traffic density estimation is proposed in this research work. The vehicular density is determined in the context of linguistic terms like low traffic, medium, and high/ heavy traffic. Mel-Frequency Cepstral Coefficient (MFCC) algorithm is modelled to extract the feature vectors for contiguous multiple frames, and classification is performed using ANFC and further performance improved using feature selection (FS) mechanism. To reduce the computational time per iteration, SSCG is employed in this research work. Approx 60 to 70% of training time is shorten per iteration.