Motivation and BackgroundPython

Java Python Motivation and Background

Consider a diKerential drive robot with wheel radius r ∈ (0, ∞), illustrated in Figure 1. Note that the wheels are perfectly aligned and placed along the robot’s y −axis, which crosses the robot centre, such that each wheel is 2/l away from the centre of the robot.

Consider that each wheel can be independently controlled with linear velocity signals vl(t) and vr(t) for the left and right wheels, respectively. The lower bound for the wheel velocities for both wheels are vmin = −0.1 and the upper bound is vmax = 0.1.

The robot is designed in such a way that when vl(t) = vr(t) > 0,  the robot moves towards its x −axis. The pose x(t) ∈ R3 of the centre of the robot with respect to the global reference frame. is such that

where px(t) is the position in the x −axis, py(t) is the position on the y −axis, and Φz is the rotation of the robot centre about the ? −axis. Note that Px(0) = py(0) = Φ(0) = 0. In addition, at x(0), the robot's reference frame. matches the global reference frame.

The sampling time of the robot is ? = 0.01. When necessary to integrate velocity into position, use the diKerence quotient equation. For instance, for any a(t) ∈ R with time derivative a(t), we have that

• All variables are defined in the International System of Units (SI), unless otherwise stated. For example, time is in seconds, lengths in meters, and angles in radians.

Figure 1 Illustration of a di2erential drive robot. The z-axis is found with the right-hand rule.

Tasks

Solve all tasks. Create a compressed file (.zip) containing the folder coursework1 containing all your submission files. The compressed file must be submitted to BB.

Task 1

I. interface_package must be a valid ament_cmake ROS2 package [1 mark]. This package is meant for semantically meaningful ROS2 interfaces used in this coursework.

II. The package must have valid ROS2 interface files in the correct folders as follows.

a. WheelVelocities.msg must have only the two fields described below [1 mark]

i. The first field must be “right_wheel_velocity” with type “float64”.

ii. The second field must be “left_wheel_velocity” with type “float64”.

b. TaskSpacePose.msg must have only the three fields described below [1 mark]

i. The first field must be “x” with type “float64”.

ii. The second field must be “y” with type “float64”.

iii. The third field must be “phi_z” with type “float64”.

c. TurnRobotOn.srv must have an empty Request and the Response should be the field “success” with type “bool”. [1 mark]

d. TurnRobotOff.srv must have an empty Request and the Response should be the field “success” with type “bool”. [1 mark]

Task total [5 marks]

Task 2

The following behaviour will only be checked for marks if all tasks above are working correctly.

I. robot_package must be a valid ament_python ROS2 Package [1 mark]

II. The package must have a valid rclpy ROS2 node with the name robot_node [1 mark]

a. The node must finish cleanly when an interrupt signal is sent and stay alive otherwise, according to the tutorials. Deviations from this might award a mark of zero for any marking below.

III. The expected behaviour of robot_node:

a. You can choose how to manage the following internal variables as long as the behaviour below is achieved.

i. The node must have an OFF state and an ON state managed by the node.

ii. The node must store internally the current pose ?(?).

iii. The values ? for the wheel radius and ? for the distance between wheels must be hard coded into the node.

b. it must have a service server providing the service robot Motivation and BackgroundPython /turn_robot_on with type TurnRobotOn.srv

i. The service must be available regardless of state. [1 mark]

ii. When the robot is in the OFF state. Upon receiving a request, it must change the robot state to ON and return “success=True”. [1 mark]

iii. When the robot is in the ON state. Upon receiving a request, it must not change the robot state and return “success=False”. [1 mark]

c. it must have a service server providing the service robot/turn_robot_off with type TurnRobotOff.srv.

i. The service must be available regardless of state. [1 mark]

ii. When the robot is in the ON state. Upon receiving a request, it must change the robot state to OFF and return “success=True”. [1 mark]

iii. When the robot is in the OFF state. Upon receiving a request, it must not change the robot state and return “success=False”. [1 mark]

d. it must have a publisher to the topic robot/task_space_pose with message type TaskSpacePose.msg. It must be used to publish the current robot pose x(t) according to the specifications below.

i. The publisher must be connected to the topic regardless of state. [1 mark]

ii. When the robot is in the OFF state. It must not publish any messages. [1 mark]

iii. When the robot is in the ON state. It must publish ?(?), that is the current pose of the robot at a frequency of 10 Hz. The fields of the message must reflect the “x”, “y”, and “phi_z” explained in Eq. (1). [1 mark]

e. It must have a subscriber to the topic robot/wheel_velocities with message type WheelVelocities.msg. It must be used to receive wheel velocity commands from an external publisher according to the specifications below.

i. The subscriber must be connected to the topic regardless of state. [1 mark]

ii. When the robot is in the OFF state. It must ignore any messages received from the topic. [1 mark]

iii. When the robot is in the ON state. The node must store the right wheel velocity and the left wheel velocity received through the topic in the internal states vr and vl, respectively. The velocities received must be clamped to be above the lower limit vmin and below the upper limit vmax. [1 mark]

f. The following behaviour will only be checked for marks if all points above are working correctly.

i. Whenever the robot is triggered to the OFF state. The internal states vr and vl must be set to zero. There should be no update of the pose x(t). [1 mark]

ii. Whenever the robot is in the ON state. The pose ?(?) must be updated at a frequency of 10 Hz according to the diGerential drive robot equations of motion. [1 mark]

Task total [15 marks]

Task 3

The following behaviour will only be checked for marks if all tasks above are working correctly.

I. controller_package must be a valid ament_python ROS2 Package. [1 mark]

II. The package must have a valid rclpy ROS2 node with the name controller_node. [1 mark]

a. The node must finish cleanly when an interrupt signal is sent or stay alive otherwise, according to the tutorials. Deviations from this might award a mark of zero for any marking below.

III. The expected behaviour of controller_node:

a. When started, it must call robot/turn_robot_on when the service becomes available. [1 mark]

b. Using ? robot/wheel_velocities, it must rotate the robot from its initial angle Φz(0) to the desired angle Φz,d given by desired_angle_deg (in degrees) in 60 seconds or less. [1 mark] Please be attentive to the conversion between degrees and radians where needed.

c. After the rotation is complete. Using robot/wheel_velocities, it must move the robot in a straight line from its initial position and Φz,d forward (about the x −axis) for a length of 1 meter (tolerance of 1 mm) in 60 seconds or less, stop, and turn oK the robot         

本指南详细阐述基于Python编程语言结合OpenCV计算机视觉库构建实时眼部状态分析系统的技术流程。该系统能够准确识别眼部区域,并对眨眼动作与持续闭眼状态进行判别。OpenCV作为功能强大的图像处理工具库,配合Python简洁的语法特性与丰富的第三方模块支持,为开发此类视觉应用提供了理想环境。 在环境配置阶段,除基础Python运行环境外,还需安装OpenCV核心模块与dlib机器学习库。dlib库内置的HOG(方向梯度直方图)特征检测算法在面部特征定位方面表现卓越。 技术实现包含以下关键环节: - 面部区域检测:采用预训练的Haar级联分类器或HOG特征检测器完成初始人脸定位,为后续眼部分析建立基础坐标系 - 眼部精确定位:基于已识别的人脸区域,运用dlib提供的面部特征点预测模型准确标定双眼位置坐标 - 眼睑轮廓分析:通过OpenCV的轮廓提取算法精确勾勒眼睑边缘形态,为状态判别提供几何特征依据 - 眨眼动作识别:通过连续帧序列分析眼睑开合度变化,建立动态阈值模型判断瞬时闭合动作 - 持续闭眼检测:设定更严格的状态持续时间与闭合程度双重标准,准确识别长时间闭眼行为 - 实时处理架构:构建视频流处理管线,通过帧捕获、特征分析、状态判断的循环流程实现实时监控 完整的技术文档应包含模块化代码实现、依赖库安装指引、参数调优指南及常见问题解决方案。示例代码需具备完整的错误处理机制与性能优化建议,涵盖图像预处理、光照补偿等实际应用中的关键技术点。 掌握该技术体系不仅有助于深入理解计算机视觉原理,更为疲劳驾驶预警、医疗监护等实际应用场景提供了可靠的技术基础。后续优化方向可包括多模态特征融合、深度学习模型集成等进阶研究领域。 资源来源于网络分享,仅用于学习交流使用,请勿用于商业,如有侵权请联系我删除!
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值