Racecar Part 3: Reflexive Control

In part 3, the goal was to create a rudimentary pipeline for following a wall while checking for collisions in front. To accomplish this, we divided the task into natural functionality segments both to better organize code structure as well as to add modularity for future functionality to build off of. In addition, we continued to add infrastructure to make writing ROS code with good practices easier.

rviz results from simulated laser scan


Table of Contents


Scan Parser

The lowest level of functionality is the scan parser node, which takes input directly from the scanner and manipulates it into a form more useful for the rest of the nodes to use. First, the laser scan is divided into 3 segments: front, left, and right. The division between the three scans is determined by input from the ROS parameter service. Each of these reduced laser scans is republished to its own topic so other nodes can look only at the portion of the scan they require, and many nodes can look at each trimmed scan without recomputing the trim operation.

The other operation the scan parser completes is to transform the LaserScans into PointCloud2s using the laser_geometry package. The point cloud format enables a node like a wall detector to access the scan information in a potentially more convenient way, for example, if it is going to perform regression or other operations on the points of the scan in cartesian space. A point cloud is published for each of the trimmed down scans as well.

The scan parser is also responsible for ignoring points which are the robot sighting itself. It does this by dropping samples outside of a particular angle band.

Object Detector

The object detector looks at the laser scan from the front of the robot to check for any close objects which could cause a collision. Using the trimmed scan from the scan parser, it counts the points within a distance threshold and publishes a boolean indicating whether there are sufficiently many points within this threshold. The threshold for distance and for the number within that distance are determined by input from the ROS parameter service.

Wall Detector

The wall detector detects the nearest point to the obstacle in trimmed scanner data from the right side, publishing the location of this point in cartesian space relative to the frame of the robot. The desired quantities of angle and range to the wall are approximated by the angle and range of this shortest beam, recoverable from the cartesian coordinates simply by converting back to polar.

We construct two of these objects using a roslaunch file containing the following:

<node pkg="reflexive_control" name="right_wall_detector" type="wall_detector.py">
  <remap from="laser/scan" to="/tokyo/laser/scan/right"/>
  <remap from="wall_detection" to="/tokyo/wall_detection/right"/>
<node pkg="reflexive_control" name="front_wall_detector" type="wall_detector.py">
  <remap from="laser/scan" to="/tokyo/laser/scan/front"/>
  <remap from="wall_detection" to="/tokyo/wall_detection/front"/>


Wall Follower

The wall follower node implements a simple two-state proportional controller to follow a wall to its right at a fixed distance. Given the input vector from the wall detector, it first extracts the distance and angle to the wall encoded in the point message. Given some target distance and angle from the wall, the node multiplies each error by a gain constant and publishes the result to the racecar’s drive system as the angle component of the ackermann command message. The target distance and angle as well as the gain constants are determined by the ROS parameter service.

Safety Controller

safety_controller_node acts as a safe and transparent way of disabling robot movement when the robot has detected that an object has encroached closer than a specified safety margin from the front of the robot. To achieve this, safety_controller_node works independently of higher-level robot logic and outputs “STOP” commands to the robot’s servos through the output MUX when it detects that an object has encroached upon the safety barrier.

To detect objects in front of the robot, the safety_controllernode subscribes to the output of the front_wall_detector node and, on every update from front_wall_detector, decides whether an object has encroached too close to the front of the robot. If an object has encroached too close to the front of the robot, the safety_controller_node will override all outgoing movement commands to the servos with a default AckermannSteeringStampedmessage that disables all driving movement by using the ackermann_cmd_mux/input/safety topic which takes precedent on all other robot actuator commands.

Safety Controller Node Hookup Diagram

Putting it all together


There are two things to avoid doing when writing ros nodes: hard-coding constants in the source code, rather than allowing them to be configured them from the launch file Naming the parameter in the source code something different from the one in rosparam Accidently forgetting to hook a subscriber up, or hooking the same handler up twice by accident

Solve this by writing a wrapper for some of rospy’s key functions, called rospyext. This used a handful of advanced python features – metaclasses, descriptors, and decorators. The end result is that instead of modules looking like this:

class ObjectDetectorNodeBefore(object):
    def __init__(self):
        self.pub_detect = Publisher('/object_detection', Bool)
        Subscriber('/laser/scan', LaserScan, self.sub_scan)
        self.distance_param = rospy.get_param('~distance_thresh', 1)

    def sub_scan(self, scan):
        detected = min(scan) < self.distance_thresh

They look like this:

class ObjectDetectorNodeAfter(Node):
    pub_detect = Publisher('/object_detection', Bool)
    distance_thresh = Param(float, default=1)

    @Subscriber('/laser/scan', LaserScan)
    def sub_scan(self, scan):
        detected = min(scan) < self.distance_thresh

Namespace Manipulation

Discovering that the topic names used by the simulator did not align with those used by the hardware. In particular many topics which are in /racecar in the simulator are in /vesc on the car. The most important of which so far is /racecar/ackermann_cmd_mux/input/*.

The approach to being able to run the same node code in the simulator as on the car with the least amount of manual tuning was to use remapping. Nodes use the simulator topic layout as canon, and a standard launch file wall_follower_sim.launchworks in the simulator. For running on the car, a launch file wall_follower_car.launch first remaps the topics  in /vescand /scan to their /racecar equivalents and then includes wall_follower_sim.launch.

Additionally, to get the simulator’s namespace to adhere as closely to the robot’s operating namespace, here I created a new Gazebo simulator launch file racecar_tunnel_with_mux.launch that also boots a ackermann_cmd_mux MUX under the /racecarnamespace. Using this modified launch file, we can now post commands to ackermann_cmd_mux/input/safety and ackermann_cmd_mux/input/navigation and expect the simulated robot to prioritize these commands in the correct manner.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s