A Self-Supervised Terrain Roughness Estimator for Off-Road Autonomous Driving

A Self-Supervised Terrain Roughness Estimator for Off-Road Autonomous Driving

David Stavens and Sebastian Thrun

Stanford Artificial Intelligence Lab

David Stavens, Sebastian Thrun

Self-Supervised Learning

“Combines” strengths of multiple sensors.

Ultra-Precise, No Range Precise, Long Range

David Stavens, Sebastian Thrun

Overview

Introduction and Motivation Classifying Terrain Roughness Self-Supervised Learning Experimental Results

David Stavens, Sebastian Thrun

2005 DARPA Grand Challenge

David Stavens, Sebastian Thrun

Velocity Planning for DGC 2005

Mobile robotics traditionally focuses on steering.

But speed is also important.– Beyond stopping distance and lateral maneuverability.

For Grand Challenge 2005, our vehicle adapted its speed to terrain conditions, minimizing shock:– Increases electrical and mechanical reliability.– Mitigates pose error for laser projection.– Increases traction for improved maneuvers.– Seems to be correlated with slowing on “hard” terrain.

David Stavens, Sebastian Thrun

Velocity Planning for DGC 2005

Simple three state algorithm:– Drive at speed limit until shock threshold exceeded.– Slow to bring the vehicle within the shock threshold.

• Uses approx. linear relationship between shock and speed.• Which is also important for the new work we present.

– Accelerate back to the speed limit.

Discontinuous control problem.– Hard to solve with conventional control approaches.

We used supervised learning.

David Stavens, Sebastian Thrun

Experiments for DGC 05

David Stavens, Sebastian Thrun

This Talk: Next Logical Step

We expand our online approach to be proactive.– Our previous approach was entirely reactive.

Difficult to be that precise with laser scanners.– Hence problems of uncertainty and learning.

Accuracy required for roughness detection exceeds that required for obstacle avoidance.– 15cm vs. 2-4cm

David Stavens, Sebastian Thrun

Other Approaches to Velocity Control

Terramechanics: guidance through rough terrain.– Online assessment only at low speeds.– High speeds require a priori maps.

Our approach is both online and at high speeds.– Speeds up to 35 mph.

David Stavens, Sebastian Thrun

CMU’s Preplanning Trailer

David Stavens, Sebastian Thrun

Overview

Introduction and Motivation Classifying Terrain Roughness Self-Supervised Learning Experimental Results

David Stavens, Sebastian Thrun

Acquiring a 3D Point Cloud

David Stavens, Sebastian Thrun

Errors in Pose and Projection

David Stavens, Sebastian Thrun

Z Error vs. Time

David Stavens, Sebastian Thrun

More than t

“Spread” of plot implies more factors than t.

t is also related to:– Amount/rate of pitching.– Distance between the two scans.

David Stavens, Sebastian Thrun

Comparing Two Laser Points

pair =1| z |2 –

3| t |4 –

5| xy distance |6 –

7| dpitch1 |8 – 7| dpitch2 |8 –

9| droll1 |10 – 9| droll2 |10

Seven Features: z, t, xy distance, dpitches, drolls 10 Parameters: 1 2 … 10 (generated with self-supervised learning)

David Stavens, Sebastian Thrun

Combining Multiple Comparisons

n pairs in ascending order.– Use weighting because resolution of discontinuities is near

resolution of laser. There are not many witness pairs.

n

R = pair 11i

i = 0

This generates a score, R, for that patch of terrain.

But how do we assign target values to R?

David Stavens, Sebastian Thrun

Overview

Introduction and Motivation Classifying Terrain Roughness Self-Supervised Learning Experimental Results

David Stavens, Sebastian Thrun

Self-Supervised Learning

Actual shock when driving over terrain modifies belief about original laser scan.

Improves classifier for subsequent scans!

David Stavens, Sebastian Thrun

Caveat: Must Correct for Speed

David Stavens, Sebastian Thrun

Mapping from R to Shock

Learn a simple suspension model in parallel with the classifier:

Rcombined = Rleft 12 + Rright

12

Rleft and Rright is for the terrain under each wheel.

David Stavens, Sebastian Thrun

Overview

Introduction and Motivation Classifying Terrain Roughness Self-Supervised Learning Experimental Results

David Stavens, Sebastian Thrun

David Stavens, Sebastian Thrun

David Stavens, Sebastian Thrun

Summary

Road shock provides ground truth for previously perceived patches of road.

Perception model improves in real-time.

Future terrain assessment is more precise.

A faster route completion time is possible.– For the same amount of shock.

Works either “offline” or “as you drive.”– Offline results presented.

David Stavens, Sebastian Thrun

Questions?