Skip to content

Latest commit

 

History

History
99 lines (90 loc) · 3.44 KB

README.md

File metadata and controls

99 lines (90 loc) · 3.44 KB
Raw

Perception Layer

基于xnxpilot的感知层模块，改写成Cyber RT中Component形式进行通信，需要在apollo项目根目录运行

Installation

third_party install

OpenCV

sudo bash docker/build/installers/install_opencv.sh

GLES3

sudo apt update
sudo apt install libgles2-mesa-dev

OpenCL

sudo apt install clinfo opencl-headers opencl-dev libpocl2

ZeroMQ

sudo apt install libzmq3-dev

libyuv

git clone https://chromium.googlesource.com/libyuv/libyuv
cd libyuv
mkdir build && cd build
cmake ..
make
make install

third_party config

OpenCV

在third_party/opencv/opencv.BUILD中，新增cc_library规则
cc_library(
name = "videoio",
includes = ["."],
linkopts = [
"-L/opt/apollo/sysroot/lib",
"-lopencv_videoio"
],
deps = [
":core",
":imgproc",
":highgui"
]

tools

在根目录下的tools/workspace.bzl文件内导入配置的第三方依赖库
新增load("//third_party/openpilot:workspace.bzl", openpilot = "repo")
然后在initialize_third_party()方法内添加openpilot()，之后就可以在BUILD文件中的deps内使用@openpilot//:repo的形式依赖自定义安装的三方库，repo换成通过cc_library自定义的name

Build project

下载该项目解压到apollo项目根目录，然后执行
bazel build perception:libperception.so
Debug模式，执行bazel build perception:libperception.so -c dbg

Run

通过cyber_launch和mainboard命令行启动各模块（camerad、calibrationd、modeld和radard）
e.g.:
cyber_launch start perception/camerad_component/camerad.launch
mainboard -d perception/camerad_compoent/camerad.dag
通过cyber_monitor监测通道数据

sequenceDiagram participant G as main participant F as radard_thread participant A as RadarD participant C as get_lead participant D as Cluster participant B as match_vision_to_cluster participant E as Track

G ->> F: main()

F ->> F: config_realtime_process
F ->> F: Ratekeeper, Priority.CTRL_LOW
F ->> F: SubMaster, PubMaster
F ->> F: RadarInterface, CarParams
F ->> F: while 1
F ->> A: update(sm, rr, enable_lead)
A ->> A: update self.current_time, self.v_ego, self.v_ego_hist
A ->> F: dat = RD.update(sm, rr, enable_lead)
A ->> pm: pm.send('radarState', dat)
A ->> pm: pm.send('liveTracks', dat)
F ->> F: rk.monitor_time()

G ->> F: main()

Note right of A: KalmanParams, Cluster, Track\nand other imports
Note right of A: laplacian_cdf, interp, math, defaultdict, deque, messaging, car, numpy_fast, Params, realtime, config_realtime_process, RADAR_TO_CAMERA\nand other imports

Note right of C: match_vision_to_cluster\nget_lead
C ->> B: match_vision_to_cluster(v_ego, lead_msg, clusters)
B ->> D: prob(c)
D ->> E: add(self.tracks[idens[idx]])
C ->> C: return lead_dict
C ->> C: low_speed_override
C ->> D: Cluster().get_RadarState_from_vision(lead_msg, v_ego)
C ->> C: lead_dict = closest_cluster.get_RadarState()
C ->> D: reset_a_lead(aLeadK, aLeadTau)
C ->> C: return lead_dict

Note right of F: enable_lead
F ->> F: can_sock, sm, pm
F ->> F: RI, RD
F ->> F: while loop, RI.update
F ->> F: RD.update
F ->> F: pm.send('radarState', dat)
F ->> F: pm.send('liveTracks', dat)

G ->> F: main()