Internet of Things(IoT)/edge intrusion-detection systems operate on distributed traffic and system-state data whose distributions vary across gateways, services, and attack conditions. We study a server-assisted federated setting in which a teacher reference is fitted from a permitted server-accessible training pool and explicitly distinguish this simulation assumption from fully decentralized deployment. The proposed framework evaluates progressive local training through boundary stabilization, confidence-weighted decision distillation, representation alignment, and validation-quality-aware aggregation. The evaluation uses a leakage-controlled protocol: server and client validation subsets are held out before federated training, update quality and early stopping use validation data only, and the final-test split is evaluated once. Results on NSL-KDD, CIC-IDS2017, Edge-IIoTset, and the ToN-IoT network dataset show competitive primary performance and stronger robustness in several severe label-skew settings. On the Telemetry of Things(ToN-IoT) with Dirichlet alpha = 0.1, the proposed method achieves 91.46 ± 5.54 F1, compared with 53.73 ± 49.00 for FedAvg and 53.77 ± 48.92 for FedProx. The results do not establish universal superiority or a universally optimal stage order: competing methods remain stronger in selected stable and attack-shift settings. The framework is therefore presented as a bounded, server-assisted robustness-oriented training strategy for heterogeneous IoT/edge intrusion detection.

Federated intrusion detection systems (FL-IDSs) for IoT and wireless sensor networks (WSNs) must remove client contributions after training when consent changes, devices retire, or compromised nodes are identified. Existing federated unlearning methods are not well aligned with transformer-based IDSs because they use coarse attribution, provide limited trajectory verification, and handle concurrent deletion requests weakly under rare-attack imbalance conditions. This paper proposes ARFU-IDS, a robust federated unlearning framework for transformer-based IoT and sensor-network IDSs. ARFU-IDS combines attention-head attribution, dual-path layer criticality probing, manipulation-resistant iterative verification, and conflict-graph scheduling to remove target-client influence while preserving retained detection utility. Experiments on UNSW-NB15, CICIoT2023, and IoTID20 evaluated detection utility, rare-category recall, adversarial robustness with network-flow feature triggers, and concurrent unlearning. On UNSW-NB15, ARFU-IDS achieves 87.1% Macro-F1 and 77.6% rare-category recall, within 0.2 percentage points of full retraining. Under flow-feature trigger attacks, it reduces attack success rate to 8.2% at $f=0.1$ and 9.7% at $f=0.2$. For three concurrent deletion requests, it shortens online unlearning latency by 43.3% relative to sequential processing. These results show that ARFU-IDS offers a practical route to robust and efficient federated unlearning for transformer-based IDSs in IoT and sensor-network deployments.