ELX304 – Electronic Systems


ELX304 – Electronic Systems 
Individual Coursework Assignment 
Digital Design 
 
SUBMISSION ONLINE on 13/10/2024 via CANVAS 
 
Introduction 
 
This coursework exercise will provide you with the opportunity to demonstrate the skills you 
have developed throughout the first half of the module. Specifically, the following learning 
outcomes will be assessed: 
1. Have a critical knowledge of the design of complex multivariable combinational and 
sequential logic circuits 
2. Have the ability to apply digital design principles to practical problems. 
This coursework contributes 50% to the overall module mark. 
 
Design Problem: 
Task 1 : Synchronous Design Problem 
The following is a model of a control system for the robot shown. The circuit inputs are the 
proximity sensor outputs (S1S2) which are a logic high when an obstacle is detected. The outputs 
(Z1Z2) are the forward drive signals to the robot’s motors where a logic high represents ‘drive’ 
and a logic low ‘brake’. 
 
Robot (viewed from above)  
 
a. Determine any operational limitations that are present in the above solution. 
 [ 5%] 
b. Design a minimal Synchronous solution to the above problem using D-type flip-flops.
 [40%] 
c. Produce a simulation of the system using suitable software. [15%] 
 
Task 2 : Asynchronous Design Problem 
A proposed new robot has only a single proximity sensor (Z) and avoids an obstacle by continuing 
to rotate in the same direction until the sensor signal is lost (Z=0). If the control strategy is to 
alternate rotation in the sequence left-right-left-etc., develop a minimal race-free asynchronous 
solution. [40%] 
 
Student Submission. 
The maximum marks obtainable for each section of the assessment are shown above, in order 
to achieve good marks, it is necessary to demonstrate an in depth understanding of the 
problem and the design procedures involved. 
You may choose your preferred method of submission, either: 
a) A full design document that shows all of the details of how your solution was developed 
and answers the questions posed. This document should be no more than 10 pages long 
and does not need long introductions or narratives. 
 
b) A video presentation where you talk through your design and answer the questions. 
This should be edited using the reView and Panopto software within Canvas. The video 
should not last longer than 10 minutes. 
 
ALL STUDENTS WILL NEED TO SUBMIT THE PROJECT SIMULATION FILE WITH THE 
CORRECT FORMAT. 
 
FURTHER INFORMATION 
Extension requests 
Requests for extensions should be directed to the module leader in the first instance. If your 
module leader is unavailable, you can contact your programme leader in their absence to request 
an extension. For further information on the regulations governing extensions of assessment 
deadlines, see here. Note that if you submit your work late without an agreed extension, this will 
be recorded as a non-submission. 
Extenuating circumstances 
In situations where a short extension may not suffice, the University of Sunderland has a 
procedure to deal with events which affect your work, but which are not predictable and are 
beyond your control, for example, illness, enabling you to defer a piece of work (and submit it at 
a later assessment point). 
Important note: You should note that the policy (like that of many universities) takes the view 
that by sitting an examination or handing in an assessment, students have deemed themselves 
fit, and no subsequent claim for extenuating circumstances will be considered. The Faculty will 
publish deadline for the submission of such claims prior to the assessment and it is important 
that you are aware that such claims must be made by the deadline, as after it has passed, a claim 
will not be considered unless there are exceptional circumstances. 
For further information on the regulations governing consideration of extenuating 
circumstances, guidance on extenuating circumstances and details of how to apply, see here 
Academic misconduct 
All work submitted is expected to be your own work. Common forms of academic misconduct 
include plagiarism and collusion, but are not limited to these. The penalties for academic 
misconduct can be very serious. If you are unsure what academic misconduct is, you should 
contact the module leader or your personal tutor to discuss it. Please familiarise yourself with 
the University’s Guide to Academic Integrity and Misconduct and the University’s Academic 
Misconduct Regulations, available here. 
 

Unet是一种深度学习模型,最初由Ronneberger等人在2015年提出,主要用于生物医学图像分割。在Matlab中实现Unet网络可以利用其强大的数学计算能力和友好的可视化界面,非常适合科研和教育用途。这个"Unet分割(Matlab)Demo"提供了在Matlab环境中构建、训练和应用Unet模型的示例。 Unet网络的特点在于其对称的架构,由下采样(编码器)和上采样(解码器)两部分组成。编码器部分用于捕捉图像的上下文信息,通过多个卷积层和池化层逐级降低特征图的分辨率,增加表示能力。解码器部分则负责恢复图像的原始空间分辨率,通过上采样和与编码器的跳连接来恢复细节信息。 在`segunet.mlx`文件中,我们可能会看到以下关键步骤的实现: 1. **网络结构定义**:定义Unet的卷积层、池化层、上采样层等。Matlab的Deep Learning Toolbox提供了构建自定义网络的函数,如`conv2d`、`maxpool2d`和`upsample2d`。 2. **损失函数选择**:图像分割通常使用交叉熵损失(cross-entropy loss),有时也会结合Dice系数或Jaccard相似度来评估模型性能。 3. **数据预处理**:`data`文件可能包含训练和验证数据,需要进行归一化、分批次等预处理操作。 4. **模型训练**:设置优化器(如Adam)、学习率策略,并执行训练循环。 5. **模型评估**:在验证集上评估模型的性能,例如计算 Dice 指数或IoU(Intersection over Union)。 6. **可视化结果**:展示模型预测的分割结果,与实际标签对比,帮助理解模型性能。 为了运行这个Demo,你需要确保安装了Matlab的Deep Learning Toolbox以及相关的数据集。`segunet.mlx`是Matlab Live Script,它将代码、注释和输出结合在一起,便于理解和执行。在Matlab环境中打开此脚本,按照指示操作即可。 此外,了解Unet在网络架构设计上的创新,比如跳跃连接(skip connections),有助于理解模型为何能有效地处理图像分割任务。Unet的成功在于它既能捕捉全局信息又能保留局部细节,因此在生物医学图像分析、遥感图像分割、语义分割等领域有广泛应用。 这个"Unet分割(Matlab)Demo"提供了一个直观的起点,帮助初学者和研究人员快速上手Unet网络的实现和训练,为后续的图像分割项目打下基础。通过学习和实践,你可以掌握深度学习在Matlab中的应用,进一步提升在图像处理领域的技能。
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