Uploaded October 2024 | Updated September 2026, 13 minutes ago

---

🔗 Download Complete Project👇
engrprogrammer-shop.fourthwall.com/products/deep-reinforcement-learningbased-bipedal-robot-walking-in-matlab-ddpg

📌instagram:
instagram.com/engrprogrammer2494

🎊Welcome to todays tech.. this video is about " Deep Reinforcement Learning Based Bipedal Robot (Actor-Critic Based) "

I'm excited to share my recent project, where I trained a 10-DOF bipedal robot to walk in a simulated environment using MATLAB SimMechanics, driven by
Deep Reinforcement Learning tool 🚀

The objective was to train the robot agent to follow these key policies:
.Move forward
.Stay on the designated path
.Avoid falling
.Torque distribution across joints
This was achieved through the reward function:
rt=vx−1.5y^2−25z^2+25Ts/Tf
−0.095∑(u^i(t-1))^2
Within 6000 training Episodes
This function fed into the neural network and optimized the robot's movements with 31 observations and 6 actions which in this state torque inputs to the robot's joints.

The results i got is not the best but very satisfied to me as it achieved the most part of the prospective results .

I’m eager to further explore and refine this exciting area of research in robotics and AI.
Looking forward to diving even deeper into this cutting-edge field!🚀


Thanks for watching
Regards
TODAYS TECH

#BipedalRobot
#DeepReinforcementLearning
#MATLAB
#SimMechanics
#AI
#Robotics
#MachineLearning
#RobotSimulation
#ReinforcementLearning
#MATLABSimulink
#AIinRobotics
#TechInnovation
#Automation
#RoboticEngineering
#ArtificialIntelligence