Welcome to Introduction to Self-Driving Cars, the first course in University of Toronto’s Self-Driving Cars Specialization. This course will introduce you to the terminology, design considerations and safety assessment of self-driving cars. By the end of this course, you will be able to: - Understand commonly used hardware used for self-driving cars - Identify the main components of the self-driving software stack - Program vehicle modelling and control - Analyze the safety frameworks and current industry practices for vehicle development For the final project in this course, you will develop control code to navigate a self-driving car around a racetrack in the CARLA simulation environment. You will construct longitudinal and lateral dynamic models for a vehicle and create controllers that regulate speed and path tracking performance using Python. You’ll test the limits of your control design and learn the challenges inherent in driving at the limit of vehicle performance. This is an advanced course, intended for learners with a background in mechanical engineering, computer and electrical engineering, or robotics. To succeed in this course, you should have programming experience in Python 3.0, familiarity with Linear Algebra (matrices, vectors, matrix multiplication, rank, Eigenvalues and vectors and inverses), Statistics (Gaussian probability distributions), Calculus and Physics (forces, moments, inertia, Newton's Laws). You will also need certain hardware and software specifications in order to effectively run the CARLA simulator: Windows 7 64-bit (or later) or Ubuntu 16.04 (or later), Quad-core Intel or AMD processor (2.5 GHz or faster), NVIDIA GeForce 470 GTX or AMD Radeon 6870 HD series card or higher, 8 GB RAM, and OpenGL 3 or greater (for Linux computers).
Lesson 4: Environment Representation
Loading...
Do curso por University of Toronto
Introduction to Self-Driving Cars
129 classificações
Na lição
Module 2: Self-Driving Hardware and Software Architectures
System architectures for self-driving vehicles are extremely diverse, as no standardized solution has yet emerged. This module describes both the hardware and software architectures commonly used and some of the tradeoffs in terms of cost, reliability, performance and complexity that constrain autonomous vehicle design.
Conheça os instrutores
Steven WaslanderAssociate Professor
Aerospace Studies
Jonathan KellyAssistant Professor
Aerospace Studies
[MUSIC]
[MUSIC]
In this lesson we'll take a closer look at the maps created to represent
the environment around our car.
I'll present an overview of the three different map types traditionally used for
autonomous driving, which you saw briefly in the previous video.
Then I'll give you a more detailed explanation about each of the maps so you
can better understand How they're created and used throughout the specialization.
The first map we will discuss is the localization map.
This map is created using a continuous set of lidar points or
camera image features as the car moves through the enviromment.
This map is then used in combination with GPS, IMU and
wheel odometry by the localization module
To accurately estimate the precise location of the vehicle at all times.
The second map is the occupancy grid map.
The occupancy grid also uses a continuous set of LIDAR points to build a map of
the environment which indicates the location of all static, or stationary,
obstacles.
This map is used to plan safe, collision-free paths for
the autonomous vehicle.
The third and final map that we'll discuss in this video is the detailed road map.
It contains detailed positions for
all regulatory elements, regulatory attributes and lane markings.
This map is used to plan a path from the current position to the final destination.
Let's take a closer look at the localization map.
As I mentioned previously,
the localization map uses recorded LIDAR points or
images, which are combined to make a point cloud representation of the environment.
As new LIDAR camera data is received it is compared to the localization map and
a measurement of the eagle vehicles position is created by aligning the new
data with the existing map.
This measurement is then combined with other sensors to estimate eagle motion and
ultimately used to control the vehicle
Here we have some recorded LIDAR data from our self-driving car.
The movement of the vehicle is clear based on the evolution of the LIDAR points
in this video.
As the car drives out of a driveway and onto the street ahead.
This detailed evolving representation of the motion of a car through its
environment Is extremely valuable for the localization module.
The localization map can be quite large, and many methods exist to compress it's
contents and keep only those features that are needed for localization.
The construction of this map will be more rigorously explained
in the next course of this specialization.
Where we discuss localization in detail.
The occupancy grid is a 2D or 3D discretized map of the static objects in
the environments surrounding the eagle vehicle.
This map is created to identify all static objects around the autonomous car,
once again, using point clouds as our input.
The objects which are classified as static include trees, buildings,
curbs, and all other nondriveable surfaces.
For example, in this grid map, if all occupied grids were colored in,
this is what the occupancy grid may look like.
As the occupancy grid only represents the static objects from the environment,
all dynamic objects must first be removed.
This is done by removing all lidar points that are found within the bounding boxes
of detected dynamic objects identified by the perception stack.
Next, static objects which will not interfere
with the vehicle are also removed.
Such as dryable service or any over hanging tree branches.
As result of these steps only the relevent writer points from static objects from
the environment remain.
The filtering process is not perfect and so
it is not possible to blindly trust the remaining points are in fact obstacles
The occupancy grid, therefore, represents the environment probabilistically,
by tracking the likelihood that a grid cells occupy over time.
This map is then relied on to create paths for the vehicle which are collusion-free.
Both the creation of this map and its application
to the local planning problem will be covered in much greater detail.
In course four of the specialization.
Let's look at an example of an occupancy grid updating over time.
Here, we see the occupancy grid visualized as the light gray square area,
under the autonomous car.
Updating the position of static objects in the environment with black squares.
As the autonomous car moves through the environment,
all stationary objects in the environment such as poles, buildings, and
parked cars, are shown as occupied grid cells.
Finally, the detailed roadmap is a map of the full road network
which can be driven by the self-driving car.
This map contains information regarding the lanes of a road,
as well as any traffic regulation elements which may affect them.
The detailed road map is used to plan a safe and
efficient path to be taken by the self-driving car.
The detailed road map can be created in one of three ways.
Fully online, fully offline,
or created offline and updated online.
A map which is created fully online relies heavily on the static object proportion
of the perception stack to accurately label and
correctly localize all relevant static objects to create the map.
This includes all lane boundaries in the current driving environment,
any regulation elements, such as traffic lights or traffic signs,
any regulation attributes of the lanes, such as right turn markings or crosswalks.
This method of map creation is rarely used
due to the complexity of creating such a map in real time.
A map which is created entirely offline is usually done by collecting data
of a given road several times.
Specialized vehicles with high accuracy sensors are driven along roadways
regularly to construct offline maps.
Once the collection is complete, the information is then labelled with the use
of a mixture of automatic labelling from static object perception and
human annotation and correction.
This method of map creation, while producing very detailed and
accurate maps, is unable to react or adapt to a changing environment.
The third method of creating detailed roadmaps is to create them offline and
then update them online with new, relevant information.
This method of map creation takes advantage of both methods,
creating a highly accurate roadmap which can be updated while driving.
In course four on motion planning, we will present a method for storing all of
the information present in a detailed roadmap called the lane length model.
Let's look at an example of a detailed roadmap.
As you can see, the lane boundaries of the detailed roadmap are visualized in red
Along with the boundaries, the center of each lane is also visualized in red.
This information is vitally important for
path-following as it provides a default path along the lane.
As you can see in this video the vehicle,
while autonomously driving, neatly follows the center of the lane.
That completes our discussion of mapping for self-driving cars.
In this video, you learned about three types of maps commonly used for autonomous
driving: The localization map, the occupancy grid, and the detailed road map.
You'll study each of these map types further as we dive into localization,
collision avoidance, and
motion planning In the remaining courses of the specialization.
Congratulations, you completed the second module
in this introduction to self-driving cars course.
In this module, you learned how to select sensing and
computing hardware in self-driving car,
how to design specific sensors [INAUDIBLE] based on the requirements of driving.
How to decompose the software system for autonomous driving.
And what the three main types of maps are that represent the environment.
In the next module, we will take a closer look at the vehicle modeling for
the purpose of precision control.
I'll see you in the next module.
[SOUND]
Explore nosso catálogo
Registre-se gratuitamente e obtenha recomendações, atualizações e ofertas personalizadas.
O Coursera proporciona acesso universal à melhor educação do mundo fazendo parcerias com as melhores universidades e organizações para oferecer cursos on-line.
© 2019 Coursera Inc. Todos os direitos reservados.
