Deep Learning on Edge: Image Courtesy: https://www.codemotion.com/

Most of the deep learning frameworks for object detection use a Feature Extractor (backbone) and an object detector (head), where head network is attached to a backbone network. Deep Learning frameworks are usually computationally intensive, consume a lot of memory and processing power. To bring these frameworks amongst the masses using edge-devices, which have constraints of memory, battery and power it is vital to have a framework which takes these parameters into account for optimum performance of low-memory devices. HarDnet (Harmonic Dense Network) is one such framework which stands out from other widely used backbone networks such as Dense FC…


Simulation of Demo Data

Co-Authors:- Makarand Mandolkar, Utkarsh Shukla

To get an idea and a link of the content here in this blog, please have a look of the previous blogs, here Blog 1 and Blog 2

The GitHub repository can be found here.

PROJECT STATUS AND TASK ACCOMPLISHED :

The more and more we were exploring, the more and more we became aware of the challenges involved in achieving the objectives defined. Nonetheless, all the predefined tasks, which are Autoware installation and its exhaustive documentation, Running Demo Data (ROSBAG) on Autoware, Implementation of Openplanner, Create and follow the waypoints…


The Goals that we try to achieve throughout the project duration are as follows:-

  1. Installation and environment setup of Autoware
  2. Running Demo Data (LIDAR and Point Cloud data) on Autoware
  3. Implementation and execution of OpenPlanner in Autoware, which will result in the execution of the following task

a) Create and follow the wavepoints in the Open-Planner b) Signal/Stop sign detection c) Lane changing d) Creating a sample vector map in Unity with stop sign and lanes and its implementation in Autoware.

4. If the time permisses, we would like to go ahead with the implementation of the A* Algorithm in…


In current time there can be seen a growing interest in self-driving vehicles. With growing computation power and connected technology, complemented with millimetre accuracy Geo-positioning systems, the tasks which earlier seemed unfathomable, now are being a reality. Many modern cars nowadays incorporate some level of smartness, thus moving towards the ultimate goal of achieving a second to none Autonomous self-driving vehicle.

But is there really a need for self-driving vehicles? Many relevant surveys suggest that Self-driving cars may revolutionize the transportation industry. The basic idea of human operator guiding a vehicle through steering and pedals are changing quickly and newer…

Rugved Hattekar

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store