Robotic bodies are inherently distributed in sensing and actuation, yet learning-based control still commonly relies on centralized information processing. This work studies the problem of information organization in communication-constrained embodied control: which computations should remain local, and which information is worth transmitting for whole-body coordination. We propose FlyCNS, an embodied information-organization framework inspired by the Drosophila brain--nerve-cord connectome. FlyCNS preserves local sensorimotor computation within each limb and enables selective long-range communication through separate ascending and descending routing pathways. From a real connectome, FlyCNS extracts the directional structural complexity of these two pathway types and uses it as a weak prior over communication allocation, while message content, transmission timing, and locomotion policies remain task-adaptive and are learned through reinforcement learning. In Unitree Go1 simulation, FlyCNS exhibits more graceful performance degradation as the communication budget is tightened. Under the most restrictive setting, it uses only about 21--22% of the communication of the full-communication reference, while still maintaining a tracking score of approximately 0.882 under both command protocols, with a gap of no more than 6.1% from the full-communication reference. These results indicate that real neural connectomes can inform not only the structural design of control networks, but also provide transferable inductive biases for information organization across embodiments, guiding robots in balancing local computation and long-range coordination under limited communication resources.
Figures & tables
Fig. 1: FlyCNS: from evolved neural organization to embodied information routing under constrained communication. The whole-CNS Drosophila connectome exhibits a hierarchical organization of information processing, in which ody-related sensorimotor processing is strongly organized within local circuits, while coordination across the body is mediated by ascending state feedback and descending modulation. FlyCNS translates this principle into a local–global control architecture for quadruped robots and uses directional structural statistics extracted from the connectome as a weak routing prior to modulate ascending and descending information flow under a shared communication budget. Specific messages, routing decisions, and motor control remain task-adaptive and are learned through reinforcement learning. Whole-CNS rendering adapted from [ 9 ]
Nominal budget
Method
Tracking S↑
Communication C↓
Pass / 3
O
S
O
S
O
S
Reference
Central
0.933±0.002
0.936±0.002
–
–
3
3
FULL
0.934±0.002
0.939±0.001
8,224
8,224
3
3
75%
ASYM
0.908±0.001
0.917±0.002
6,170
6,169
3
3
PERIODIC
0.916±0.001
0.925±0.003
6,169
6,169
3
3
FlyCNS
0.926±0.003
0.930±0.004
5,480
5,278
3
3
TABLE I: Final control–communication results. O/S denote official/scripted commands; S is the tracking score and C is realized evaluation communication in logical bits per control step. Scores are mean ± sample SD across three development training seeds. “Pass” counts seeds meeting the numerical locomotion criteria in each mode, not episode survival.
Budget
Method
Mode
Rv↑
NS (%) ↓
ev↓
eω↓
C↑↓
C↓↓
75%
ASYM
O
0.915
0.072
0.162
0.109
3,077
3,093
S
0.909
0.100
0.136
0.104
3,077
3,093
PERIODIC
O
0.914
0.055
0.157
0.102
3,077
3,093
S
0.906
0.217
0.130
0.094
3,077
3,093
FlyCNS
O
0.930
0.047
0.145
0.096
2,548
2,932
S
0.923
0.183
0.125
0.092
2,465
2,813
TABLE II: Motion quality and directional communication statistics across communication budgets. Entries are means over three training seeds; O/S denote official/scripted commands. Rv is realized/commanded planar speed, NS is the near-static percentage, and ev and eω are RMSE in m/s and rad/s. C↑ and C↓ denote measured ascending and descending logical communication in bits per control step.
Fig. 3: Comparison under the 25% nominal budget. Columns correspond to the forward, lateral, yaw, and combined forward/reverse-yaw command phases at t=3,8,13,18s . Dashed black curves indicate command-integrated reference trajectories, colored curves show the actual trajectories, and arrows indicate robot heading. Insets show the cumulative motion paths.
Fig. 4: Learning dynamics for configurations targeting 25% communication. Top labels indicate the FlyCNS training curriculum; every policy is evaluated at the final 25% communication target.
While deep learning models achieve state-of-the-art performance in complex tasks, they remain brittle when faced with new environments or sensory deprivation. In contrast, biological systems exhibit remarkable tolerance to these challenges. We address this vulnerability by developing a recurrent neural network (RNN) whose architecture is directly derived from the synaptic-resolution brain connectome of the fruit fly Drosophila melanogaster. We demonstrate the feasibility of training the fly connectome neural network (FLYNN) to perform vision-based navigation in MuJoCo, achieving performance comparable to modern hand-crafted networks of similar parameter counts. Crucially, FLYNN exhibits superior resistance to out-of-distribution (OOD) data and tolerance to sensory loss without further training. It remained functional even under total vision loss while hand-crafted networks largely failed, even when specifically trained with camera dropout. Principal Component Analysis (PCA) of the internal state of FLYNN suggests that it exhibits a particularly high degree of representational modularity, which might be related to its robustness. Our work provides a new direction for designing resilient artificial agents following the topology of biological brains.
Generalized morphology control requires a single policy to transform information across limbs with different physical roles, coordinate whole-body motion, and remain efficient as body size grows. Existing communication mechanisms address these requirements only partially. We introduce RecMorph, a topology-guided spatial recurrent architecture that uses recurrent sequence computation to jointly perform cross-limb communication and representation transformation. A depth-first traversal converts the kinematic tree into a morphology-derived sequence, along which shared bidirectional transitions progressively transform limb information before action decoding. Residual preservation, RMS normalization, and input-dependent channel modulation stabilize this repeated spatial transformation, yielding linear token complexity at fixed model width and depth. Across five UNIMAL tasks, RecMorph achieves the strongest mean final training performance among the evaluated generalized morphology controllers and the highest measured inference throughput on FT, while generalizing to unseen variations and bodies with up to 30 limbs. We further migrate representative generalized controllers from UNIMAL benchmarks to a four-platform quadruped setting. RecMorph achieves the best macro-averaged performance under nominal and high friction, reduces nominal velocity RMSE by 43.5% relative to specialist MLPs, and one shared policy completes 40 physical Go1/Go2 trials without falls. These results show that topology-guided recurrent transformation provides an effective and efficient communication mechanism for Generalized Morphology Control and remains effective when transferred from procedural bodies to physical robot platforms. Code and experimental resources are publicly available at https://github.com/quanruirao/RecMorph.
Deploying learned multi-robot models on heterogeneous robots remains challenging due to hardware heterogeneity, communication constraints, and the lack of a unified execution stack. This paper presents NeuroMesh, a multi-domain, cross-platform, and modular decentralized neural inference framework that standardizes observation encoding, message passing, aggregation, and task decoding in a unified pipeline. NeuroMesh combines a dual-aggregation paradigm for reduction- and broadcast-based information fusion with a parallelized architecture that decouples cycle time from end-to-end latency. Our high-performance C++ implementation leverages Zenoh for inter-robot communication and supports hybrid GPU/CPU inference. We validate NeuroMesh on a heterogeneous team of aerial and ground robots across collaborative perception, decentralized control, and task assignment, demonstrating robust operation across diverse task structures and payload sizes. We plan to release NeuroMesh as an open-source framework to the community.