Bionik: Yeah, a swarm with a heat tracker is a solid start, but jungle acoustics are a nightmare. A predictive AI layer could learn the jungle noise pattern and filter it out, but you’ll need a lot of labeled data to train it. Also, keep the algorithm lightweight—those drones don’t have GPUs for a full neural net. Maybe start with a simple Kalman filter that predicts the ghoul’s path and then overlay a small CNN to refine the heat signal. Test it in a controlled environment first before deploying in the field.