Evaluate Localization Accuracy and Timing Under Faults and Overload

未关闭
#765 0 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看

维护者通常 2 天内回复

还没有人认领这个 Issue。

评估

难度
5/5
预计耗时
一周以上
新手友好度
12/100
Issue 类型
功能
描述清晰度
需要澄清
活跃度
活跃
技术栈
rust

调研方向

The issue names no source files or tests. Start by locating the M2 instrumentation and the input and overload policy definitions it refers to, then the localization pipeline entry point. Done means fault-free baseline and fault scenario runs that report accuracy, latency, freshness, deadline misses and data loss, with machine-readable results and reproduction commands.

由索引模型根据 Issue 内容生成。

描述

Evaluate the integrated Awkernel localization pipeline under controlled input faults and computational overload. Reuse the M2 instrumentation to quantify localization accuracy, end-to-end latency, input data age, deadline misses, and dropped-data rates, including recovery after abnormal conditions end.

Scope:

  • Establish a fault-free baseline with motion distortion correction enabled
  • Run documented scenarios involving delivery delays, timestamp offsets, message loss, and computational overload
  • Evaluate individual faults and selected combinations using the defined input and overload policies
  • Measure accuracy, latency, freshness, deadline behavior, and data loss throughout each run
  • Analyze recovery behavior and accuracy–timing trade-offs

Acceptance criteria:

  • Each experiment records its dataset, ground-truth source where available, configuration, fault schedule, workload, and policy settings.
  • Position and orientation errors are measured against aligned ground truth; reference-output differences are reported separately.
  • Timing summaries include median, tail percentiles, maximum observed values, and sample counts.
  • Input age is reported per relevant stream; timestamp-offset experiments distinguish apparent age from age derived from the preserved original timeline.
  • Deadline-miss ratios use documented job populations, with unfinished or cancelled jobs accounted for explicitly.
  • Dropped-data rates specify their denominator and distinguish injected loss, buffer overflow, stale-input rejection, and discarded late results without double counting.
  • Missing outputs and prediction-only intervals remain visible rather than being excluded from accuracy and latency summaries.
  • Recovery time and any persistent degradation after fault removal are reported using predefined criteria.
  • Repeated runs quantify variability, and machine-readable results and reproduction commands are retained.

Out of scope:

New instrumentation or fault-injection frameworks, changes to localization algorithms or scheduling policies, and formal safety or worst-case timing guarantees.

主要语言
Rust
星标
282
派生
12
平均合并
2 天 9 小时
30 天内合并 PR
28

环境准备

这个项目没有提供开发容器、Dockerfile 或贡献指南,环境需要你自己搭建:先看它的 README,通用步骤见我们的新手贡献指南。

从这里开始

  1. 先读完整个 Issue,再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库,在一个分支上完成修改。
  4. 提交 Pull Request,并在描述里引用这个 Issue 编号。

tier4/awkernel 的其他 Issue

查看 tier4/awkernel 的全部 Issue

相似的 Issue

更多 Rust Issue

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。