B2X Networks: Joint Design of Communication and Control for Embodied Intelligence

Abstract: This article proposes the concept of \emph{brain-body-to-everything (B2X)} networks to facilitate the integration of wireless networks and embodied intelligence. In this framework, the \emph{brain} refers to the intelligence functions for reasoning, planning, and decision-making, the \emph{body} denotes the physical embodied agent that senses and acts in the real world, and \emph{X} represents the surrounding ecosystem involved in the brain-body interaction loop. Two B2X architectures with \emph{distributed} and \emph{centralized} brains are introduced to characterize different placements of intelligence across the body, base station, and core network. The uplink and downlink designs of B2X networks are then discussed under a representative base-station-side brain setting. For the uplink, communication is redesigned for B2X state acquisition under event urgency, sensing volume, and simultaneous multi-body access. For the downlink, communication is redesigned to coordinate command delivery and conventional service under shared radio resources. Based on these uplink and downlink considerations, a communication-control Pareto boundary is further used to characterize the loop-level trade-off between wireless transmission performance and control quality in B2X networks. Finally, several open research problems are discussed to guide future B2X network design.

• The paper introduces the B2X framework that jointly optimizes wireless communication and control, establishing a closed-loop system for embodied intelligence.
• It details two architectures—distributed and centralized brains—and novel uplink/downlink regimes to meet stringent low-latency and reliability requirements.
• The study characterizes a communication-control Pareto frontier, balancing delay, reliability, and control error for optimized performance in autonomous systems.

B2X Networks: Joint Design of Communication and Control for Embodied Intelligence

Introduction

The paper "B2X Networks: Joint Design of Communication and Control for Embodied Intelligence" (2607.00537) rigorously develops the conceptual and architectural foundations for integrating wireless communication systems with embodied intelligence, encapsulated in the brain-body-to-everything (B2X) framework. The B2X paradigm is motivated by the shift from purely digital AI systems to intelligent agents that interact, perceive, and act within the physical world, imposing stringent and qualitatively novel requirements on wireless infrastructure. The study targets experienced researchers in wireless, control, and embodied AI, emphasizing joint design of communication and control (JDCC) as a necessary departure from conventional siloed approaches.

B2X Networks: Conceptual Foundations and Architectures

B2X networks are defined as closed-loop systems that interconnect “brain” (intelligence, reasoning, and planning), “body” (physical agent acting and sensing), and “X” (external ecosystem comprising other agents, infrastructure, and context) via wireless infrastructure. Distinct from standard wireless systems focusing on throughput and latency, B2X frames the wireless system as an intelligence-bearing infrastructure, supporting embodied agents that require context-aware, low-latency, and robust bidirectional communication loops.

The paper differentiates two primary B2X architectures:

• Distributed Brain: Intelligence is partially colocated at the body, base station (BS), and core network, reflecting operational time scales and reliability requirements. Local, BS-edge, and core intelligence jointly execute tasks ranging from millisecond safety controls to long-term, city-scale orchestration. This approach exploits spatial and temporal task diversity but requires mechanisms for authority management, state consistency, and real-time synchronization under channel variability and delay constraints.
• Centralized Brain: Intelligence is concentrated either on the body, at the BS, or within the core network. The BS-side centralized brain is emphasized as a tractable and impactful setting for analysis, striking a balance between compute proximity (for low-latency inference and control) and global context (via infrastructure-supported aggregation and planning).

This architectural taxonomy allows for precise placement of intelligence as a function of task requirements, channel conditions, and system resources, establishing brain placement as a control-design variable.

Uplink and Downlink Communications in B2X

Uplink: State Acquisition

Unlike traditional uplink design focused on rate and reliability, B2X uplink prioritizes the acquisition of control-relevant state, accounting for urgency, information content, and physical dynamics. Three operational regimes are identified:

• Event-triggered Low-latency Access (ETLA): For rare but critical events demanding immediate attention. Schemes such as event-driven reporting and grant-free access prioritize minimal reaction latency over volume.
• High-volume Sensing Access (HVSA): Supports continuous streams of rich multimodal data (e.g., LiDAR, video). Communication burden is mitigated via task-oriented compression, semantic feature extraction, and inference-split transmission.
• Massive Short-packet Access (MSPA): In multi-agent settings, large cohorts send brief but timely status/data updates. Scalable mechanisms (over-the-air aggregation, age-aware scheduling) preserve system-level timeliness and robustness.

The overall objective is to convey sufficient, fresh, and relevant information to maximize state-estimation and decision quality at the BS-side brain, subject to stringent resource constraints.

Downlink: Command Delivery

The downlink is responsible for communicating intelligence-driven commands and coordination signals back to the physical agents. The paper specifies three principal regimes:

• Data-command Multiplexed Delivery (DCMD): Joint scheduling of high-volume data and delay-sensitive control commands across shared radio resources. Priority and segmentation schemes ensure URLLC requirements are not compromised by eMBB traffic.
• Periodic Command Delivery (PCD): For tasks necessitating regular control signals, mechanisms emphasize low delay jitter and consistent command cadence.
• Mobility-robust Command Delivery (MRCD): Maintains command reliability in the presence of rapid channel dynamics and physical agent mobility, leveraging techniques such as adaptive beam management and priority-based resource allocation.

The analysis shifts the evaluation criterion from generic throughput to execution-criticality—the impact of communication performance on physical action outcomes.

Communication-Control Pareto Boundary

A central theoretical construct introduced is the communication-control Pareto boundary, which formalizes the trade-off between communication resources (state acquisition delay, command reliability) and control system metrics (steady-state trajectory error, task-level performance). This loop-level analysis transcends conventional link-level optimizations by capturing the direct and coupled influence of communication design on overall control objectives.

For example, in UAV trajectory control or robotic manipulation, the trade-off is characterized in the delay-error plane: decreasing uplink delay (improving state freshness) typically increases bandwidth/resource consumption, but in turn can decrease control error; yet, excessive resource dedication in one direction negatively impacts the other. The paper demonstrates that achievable operating points lie on a Pareto frontier, which guides protocol and resource design not by isolated communication or control maximization, but by holistic loop-level system objectives (Gan et al., 9 Apr 2026).

Open Problems and Directions for Future Research

Several critical research directions are identified:

• Fundamental Loop-level Theory: Development of analytical tools linking wireless delay, reliability, and resource metrics to tracking accuracy, stability margins, and safety in embodied control loops.
• Joint Protocol and Resource Optimization: Function- and energy-aware protocol co-design for real-world feasibility, particularly under battery constraints. This includes access, scheduling, beamforming, and edge offloading, optimized for task-driven system metrics.
• Scalable Distributed Intelligence: Algorithms and architectures for state/authority consistency, model synchronization, and robust control across distributed-brain settings with dynamic agent populations.
• Deployment and Adaptivity: Mechanisms for brain placement and migration as a function of task urgency, system reliability, and environmental context.

Practical and Theoretical Implications

The B2X framework provides a rigorous foundation for future AI-native network design, where communication and control performance are inseparable. This has clear implications for urban autonomous transportation, collaborative robotics, smart manufacturing, and large-scale UAV deployments, enabling new forms of intelligence distribution and interaction previously unsupported by existing wireless systems.

Theoretically, the communication-control Pareto frontier demands a shift in both network and control system design theory, where protocols, resource allocation, and agent policies are explicitly co-optimized under joint loop-level models. This convergence points to JDCC as an indispensable principle in embodied AI networking.

Conclusion

The B2X framework systematically extends the boundaries of wireless network design into the domain of embodied intelligence, emphasizing closed-loop co-design over independent communication or control optimization. Through its architectural taxonomy, loop-level trade-off characterization, and identification of open problems, the paper (2607.00537) sets the stage for next-generation JDCC research and practical engineered systems in AI-RANs, autonomous systems, and networked physical intelligence.