Lane Keeping with AI - Versionsgeschichte

Moye-nyuysoni.glein-perry@stud.hshl.de: /* Simulation Output */

2025-04-16T07:48:06Z

Simulation Output

← Nächstältere Version		Version vom 16. April 2025, 07:48 Uhr
Zeile 678:		Zeile 678:
	===Simulation Output===		===Simulation Output===

	[[Datei:SimulationLK.mp4\|500px\|mini\|Simulation Lane keeping]]		[[Datei:SimulationLK.mp4\|500px\|mini\|Fig. 5: Simulation Lane keeping]]

	= Task =		= Task =

2025-04-16T07:47:14Z

Results

← Nächstältere Version		Version vom 16. April 2025, 07:47 Uhr
Zeile 660:		Zeile 660:

	= Results =		= Results =
			The simulation output, which visually overlays the predicted steering direction on the original test video, provided immediate visual feedback on the system's performance. For each frame:

			* A blue indicator marked the center of the frame.

			* A red arrow, whose direction and length scaled with the predicted steering angle, graphically demonstrated the intended steering correction.

			* A side-by-side plot compared the smoothed predicted steering angles against the actual normalized angles over the sequence of frames
			== Summary of Findings ==
			* The neural network is capable of reliably predicting steering angles that largely agree with the actual values observed from the test video.

			* Data normalization and preprocessing steps were crucial in achieving consistency between the training and testing environments.

			* The simulation visuals corroborate the numerical evaluation, showing that the predicted steering corrections are both accurate and smoothly transitioning over time.

			* The results offer a promising outlook for further enhancements, such as incorporating more diverse data augmentation techniques or fine-tuning the network architecture to improve performance further.

	===Simulation Output===		===Simulation Output===

2025-04-16T07:40:26Z

Testing the Network

2025-04-16T07:32:45Z

Results

← Nächstältere Version		Version vom 16. April 2025, 07:32 Uhr
Zeile 387:		Zeile 387:
	Additionally, a PID controller is implemented to smooth the output of the network, ensuring that the control signals are both responsive and stable.		Additionally, a PID controller is implemented to smooth the output of the network, ensuring that the control signals are both responsive and stable.

			== Testing the Network ==
	= Results =		= Results =

2025-04-16T07:32:04Z

Implementation in MATLAB®

← Nächstältere Version		Version vom 16. April 2025, 07:32 Uhr
Zeile 282:		Zeile 282:
	fullyConnectedLayer(1) % Regression output		fullyConnectedLayer(1) % Regression output
	regressionLayer		regressionLayer
	];		];

	Purpose:		Purpose:
	Builds a deep neural network for processing the 64×64 grayscale images.		Builds a deep neural network for processing the 64×64 grayscale images.

2025-04-16T07:27:12Z

Visualized Data of Sample02

← Nächstältere Version		Version vom 16. April 2025, 07:27 Uhr
Zeile 28:		Zeile 28:
	=== Visualized Data of Sample02 ===		=== Visualized Data of Sample02 ===
	[[Datei:Rundkurs04.avi.mp4\|300px\|frameless\|centered\|alternativtext=Rundkurs03 output video \|Rundkurs03 ]]		[[Datei:Rundkurs04.avi.mp4\|300px\|frameless\|centered\|alternativtext=Rundkurs03 output video \|Rundkurs03 ]]

			Fig 2. Sample Input Data

	== System Requirements ==		== System Requirements ==

2025-04-16T07:26:44Z

Visualized Data of Sample01

← Nächstältere Version		Version vom 16. April 2025, 07:26 Uhr
Zeile 24:		Zeile 24:
	Lane-keeping refers to the ability of a vehicle to detect and maintain its position within a designated lane on the road. It is a critical function for autonomous driving, helping to enhance road safety and reduce driver workload. Our target location is the mechatronics lab at HSHl. below are videos recorded using videos from a [[JetRacer: Autonomous Driving, Obstacle Detection and Avoidance using AI with MATLAB\|JetRacer]]		Lane-keeping refers to the ability of a vehicle to detect and maintain its position within a designated lane on the road. It is a critical function for autonomous driving, helping to enhance road safety and reduce driver workload. Our target location is the mechatronics lab at HSHl. below are videos recorded using videos from a [[JetRacer: Autonomous Driving, Obstacle Detection and Avoidance using AI with MATLAB\|JetRacer]]
	=== Visualized Data of Sample01 ===		=== Visualized Data of Sample01 ===
	[[Datei:Rundkurs03.avi.mp4\|300px\|~~frameless~~\|centered\|alternativtext=Rundkurs03 output video \|Rundkurs03 ]]		[[Datei:Rundkurs03.avi.mp4\|300px\|Frameless\|centered\|alternativtext=Rundkurs03 output video \|Rundkurs03 ]]

	=== Visualized Data of Sample02 ===		=== Visualized Data of Sample02 ===

2025-04-16T07:24:50Z

Implementation in MATLAB®

← Nächstältere Version		Version vom 16. April 2025, 07:24 Uhr
Zeile 160:		Zeile 160:

	Converting to grayscale and resizing reduces the computational load. Adjusting contrast can highlight essential features in the images.		Converting to grayscale and resizing reduces the computational load. Adjusting contrast can highlight essential features in the images.
			[[Datei:Sample Image(SP).png\|500px\|mini\|Fig 4: Sample Training Data Generated]]

	'''Data Augmentation''':		'''Data Augmentation''':

2025-04-16T07:18:04Z

CNN from Scratch

2025-04-16T07:10:18Z

Research and Methodological Approach

@@ Zeile 47: / Zeile 47: @@
 = Research and Methodological Approach=
 === Solution Alternatives and Morphological Box===
 The morphological box is used to systematically explore and compare various solutions based on design parameters and their possible values. The following table summarizes the alternatives: