Artificial Intelligence II

3:00 PM – Deep Learning-Based Modulation Classification Using Synthetic and Over-the-Air SDR Signals
Hovannes Kulhandjian and Elizabeth Batz, California State University, Fresno; Michel Kulhandjian, Rice University

This work presents a convolutional neural network (CNN) for automatic classification of digital modulation schemes using signals received via software-defined radios (SDRs). A synthetic dataset was created in MATLAB for BPSK, 8-PSK, 16-PSK, QAM, 16-QAM, and 64-QAM, with 1,500 messages per scheme at five SNR levels and added phase noise. A 12-layer CNN trained on this dataset achieved 97% accuracy in classifying modulation types. To evaluate real-world performance, over-the-air signals were captured and used for validation, yielding classification accuracies ranging from 72% to 91%. While performance on live signals showed variability, the results indicate strong potential for generalization with further refinement. Enhancing the synthetic dataset with additional channel impairments may improve model robustness and real-world applicability. This research demonstrates the viability of using deep learning for signal classification in intelligent communication systems.

3:20 PM – Wideband Microphone Array System for Sound Localization and Species Identification in Wind Farm Environments
Lizette Crooks, Morgan Parslow, Derek Dela Cruz, Tien Nguyen, Max Stewart, and Yogananda Isukapalli, University of California, Santa Barbara

Wind energy facilities provide a sustainable power source but pose a significant threat to birds and bats that traverse these areas. To mitigate this risk, an acoustic monitoring system was designed to track and classify volant species. The system consists of an array of wideband microphones that enable time difference of arrival (TDOA) based 3D source localization using the generalized cross-correlation phase transform (GCC-PHAT) algorithm, accurate to within 1 degree in isolated environments. Furthermore, the microphones utilize a convolutional neural network (CNN) for species identification. The collected data maps migration patterns in the areas surrounding wind energy facilities, providing wind energy companies with actionable insights to adjust operations and reduce harm to wildlife.

3:40 PM- Real-Time Model-Based FTI Integrity Monitoring with AI
Dr. Nelson Paiva Oliveira Leite, MsC. Lucas Benedito dos Reis Sousa, Wagner de Oliveira Lima, Instituto de Pesquisas e Ensaios em Voo (IPEV)

FTI migrated from an architecture based on a PCM DAS to a networked distributed system, whose central element is now a Network File Server. FTI complexity and parameter count has increased significantly, encompassing the acquisition of a huge number of measurements. So, in several cases nobody knows whereas the parameter gathering is good or not. IPEV has successfully developed a Flight Test Simulator (FTS) to improve both test flight safety and test pilot and engineer formation syllabus. To improve aircraft model accuracy IPEV is now connecting into the Real-Time environment, the test bed, the Ground Telemetry System (GTS) and FTS, so parameter identification process and model tunning could be executed while the aircraft is flying. When the simulation model becomes accurate, we will be able to integrate an AI-based background process to monitor the integrity of parameters that are not observed by the ground crew. This paper discusses IPEV actual and future efforts to achieve such goal.

4:00 PM-4:40pm – Introduction to Artificial Intelligence and Machine Learning (AI/ML)
Dr. Bob Touchton, Advanced Autonomy Specialists, Inc.

This presentation is aimed at those who are perhaps new to AI/ML, yet lead, work, or communicate with AI/ML practitioners, or just want to get their arms around this rapidly growing technology. Based on my 40 years of helping organizations incorporate AI/ML to solve real world problems I will walk through what AI/ML is, and isn’t, and what capabilities distinguish a system as being “intelligent.” The history, role and challenges of AI/ML will be explored and the lingo and acronyms surrounding the 4 types of AI/ML will be introduced. Participants leave with an expanded vocabulary and broader understanding of AI/ML and an appreciation of the benefits available, and the hurdles faced by the adoption of intelligent systems.