Automation Workflow for Joint Communication and Sensing (JCAS) Resource Allocation

Joint Communication and Sensing (JCAS) networks represent a fundamental shift, using the same radio hardware and spectrum for both user data transmission and environmental sensing, such as traffic monitoring or intrusion detection. The core operational bottleneck is the manual, static scheduling of these competing functions, which leads to inefficient resource use, missed sensing opportunities, or degraded communication KPIs. A custom automation workflow eliminates this by implementing real-time, policy-driven orchestration that continuously evaluates priority, demand, and regulatory constraints to maximize asset utilization and service revenue.

Implementation requires integrating the orchestrator with the RAN Intelligent Controller (RIC) via standardized O-RAN interfaces (E2, A1) to enact scheduling decisions, and with a network digital twin for safe simulation of resource trade-offs. Critical controls include human-in-the-loop approval gates for policy changes, real-time observability dashboards for both communication and sensing performance, and robust exception routing for conflict resolution. This architecture transforms JCAS from a theoretical concept into a production system that creates new service paradigms while protecting existing communication SLAs.

This workflow automates the novel scheduling challenge of Joint Communication and Sensing (JCAS), where a single radio network must serve both user traffic and environmental sensing (e.g., traffic monitoring, intrusion detection). It eliminates the manual, rule-based partitioning of spectrum, replacing it with a real-time, multi-agent system that continuously evaluates priority, interference, and SLA constraints. The operational upside comes from maximizing the dual-use value of capital-intensive RAN assets, creating new sensing-as-a-service revenue streams while protecting core communication KPIs.

Implementation integrates with the RAN Intelligent Controller (RIC) via standardized O-RAN interfaces (A1, E2) for real-time control. The orchestrator, built on frameworks like LangGraph, manages specialized agents for communication forecasting and sensing analysis. Critical controls include policy-based guardrails to prevent sensing from violating communication SLAs, a human-in-the-loop approval gate for major resource shifts, and comprehensive observability dashboards tracking utilization, interference, and service fulfillment. Rollout requires phased deployment, starting with non-critical spectrum bands and expanding as model confidence is validated.

This blueprint details a three-phase integration to automate JCAS resource allocation, directly targeting the operational cost of manual spectrum planning and the revenue upside from new sensing services. Phase 1 establishes a non-invasive digital twin for safe experimentation, using live traffic and geolocation feeds to train scheduling agents without impacting production KPIs. This sandboxed environment validates the AI's ability to dynamically partition time-frequency resources between communication throughput and sensing resolution, building confidence in the control logic and its integration with the RAN Intelligent Controller (RIC) and OSS.

Phase 2 implements a human-in-the-loop control system on a non-critical pilot cell cluster, where the AI scheduler's allocation decisions are executed only after network engineer approval. This stage integrates the workflow with the element management system (EMS) for configuration pushes and establishes observability dashboards for KPI impact. The final production phase scales the autonomous scheduler across the network, enabling full closed-loop control with automated rollback triggers and detailed audit trails for compliance, ultimately delivering the targeted 15-25% improvement in spectral efficiency and enabling new sensing-as-a-service revenue streams.

A JCAS resource allocation workflow automates the novel scheduling challenge of partitioning time-frequency resources between user data traffic and sensing functions like traffic monitoring or occupancy detection. The operational upside comes from monetizing sensing as a new service while protecting communication SLAs, creating a dual-revenue stream from the same infrastructure. Implementation requires a policy-driven orchestrator integrated with the RAN Intelligent Controller (RIC) and OSS to execute low-latency, millisecond-scale scheduling decisions based on priority queues and interference models.

Rollout requires phased deployment with strict controls, starting in non-critical geographic areas. The orchestrator must integrate with service management and billing systems to track resource consumption for chargeback. Governance is enforced through a human review layer that audits allocation logs, intervenes on policy exceptions, and approves sensing service activation requests. Monitoring via the RIC and a digital twin provides observability into interference and SLA compliance, ensuring the autonomous system operates within defined operational and regulatory boundaries.

The primary constraint is maintaining communication service-level agreements (SLAs) while opportunistically allocating time-frequency resources for sensing. The workflow's orchestrator must continuously validate that sensing beam patterns and power levels do not create harmful interference or violate regulatory masks. This is enforced by a policy engine that ingests real-time KPI feeds from the RAN Intelligent Controller (RIC) and OSS, triggering immediate resource reclamation if communication latency or packet loss thresholds are breached. The business value is enabling new sensing revenue without risking core connectivity SLAs or compliance.

Exception handling is multi-tiered. For minor SLA drifts, the system autonomously applies compensating adjustments, like reducing sensing bandwidth. For critical breaches or policy conflicts, it escalates to a human engineer in the NOC via a dedicated dashboard, presenting the conflict, proposed resolution, and audit trail. All decisions and sensor data outputs are logged with tamper-evident timestamps for regulatory compliance. Implementation integrates with ServiceNow for ticketing and Splunk for observability, ensuring the workflow is both resilient and auditable within existing telecom operations frameworks.