Spectral and Conformal: Neural Network Methods for Calibrated Trajectory Prediction and Control-Point Intervals
Spectral and Conformal: Neural Network Methods for Calibrated Trajectory Prediction and Control-Point Intervals
This research introduces a framework utilizing neural networks with spectral backbones and modified conformal quantile regression to provide calibrated trajectory prediction and control-point intervals. The method integrates interval arithmetic to extend noise-free coverage guarantees, achieving an inference time of approximately 100 Hz (9.81 ms total pipeline) on a Raspberry Pi Zero 2 W for a quadrotor under closed-loop linear quadratic regulation.
The authors present a neural network framework combining spectral backbones and conformal quantile regression to achieve calibrated trajectory prediction and control-point intervals for autonomous navigation, reporting an inference time of approximately 100 Hz on a Raspberry Pi Zero 2 W under specific test conditions.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
inference_time_ms
3.45
View original evidence
under a closed-loop linear quadratic regulator is 3.45~ms on a Raspberry Pi Zero 2 W. The total pipeline time, including trajectory inference time, is 9.81~ms, or approximately 100~Hz, consistent with typical control rates.
Open source S4pipeline_time_ms
9.81
View original evidence
under a closed-loop linear quadratic regulator is 3.45~ms on a Raspberry Pi Zero 2 W. The total pipeline time, including trajectory inference time, is 9.81~ms, or approximately 100~Hz, consistent with typical control rates.
Open source S4frequency_hz
≈100
View original evidence
under a closed-loop linear quadratic regulator is 3.45~ms on a Raspberry Pi Zero 2 W. The total pipeline time, including trajectory inference time, is 9.81~ms, or approximately 100~Hz, consistent with typical control rates.
Open source S4
- paper: not_reported
- code: not_reported
- weights: not_reported
- dataset: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
ural network with a spectral backbone to learn trajectories and quantile bands with built-in constraints (initial conditions and stability); (2) a modified conformal quantile regression to provide calibrated, joint trajectory-level coverage; and (3) interval arithmetic to extend noise-free coverage guarantees to control points. After refactoring the network for inference, the trajectory inference time for a quadrotor
Open source S3
under a closed-loop linear quadratic regulator is 3.45~ms on a Raspberry Pi Zero 2 W. The total pipeline time, including trajectory inference time, is 9.81~ms, or approximately 100~Hz, consistent with typical control rates.
Open source S4
Source：arXiv Robotics — research abstracts · arxiv.org