brianletort.ai
← Library

NASA JPL

Verified evidenceCreator to judge

Perseverance Enhanced Autonomous Navigation

Communication latency requires local path authority.

Engineering design · Mars

Collections: Human still decides · Embodied work

An editorial scene for NASA JPL contrasts use orbital and rover imagery to define route goals and commanded drive segments. with builds local terrain maps, checks collision/slip risk, and selects paths onboard. in the perseverance enhanced autonomous navigation workflow.

Executive brief

The operating-model shift, in one view.

Humans specify intent; the machine owns safe local action.

AI value · Autonomous distance share

Verified

About 90% of 32.1 km used ENav

Cross-generation mission comparison.

Before

Use orbital and rover imagery to define route goals and commanded drive segments. → Executes a commanded segment and waits for telemetry review before subsequent planning.

After

Builds local terrain maps, checks collision/slip risk, and selects paths onboard. → Sets goals and corridors, reviews telemetry, and handles hazards that stop autonomous progress.

Human boundary

ENav owns local motion; Earth owns mission goals.

Why it matters

Communication latency requires local path authority.

How the work changed

Before

How the work ran before the change.

  1. Step 1 of 2

    Earth rover planners

    Use orbital and rover imagery to define route goals and commanded drive segments.

    ControlCommunication delay prevents real-time joystick control; prior rover comparison provides no uniform terrain-adjusted cycle time.

  2. Step 2 of 2

    Rover and ground team

    Executes a commanded segment and waits for telemetry review before subsequent planning.

    ControlGround mission safety review.

What changed

Communication latency requires local path authority.

Decision rightHuman moves from creator to judge

After

How the same work runs now.

  1. Step 1 of 2

    Enhanced AutoNav

    Builds local terrain maps, checks collision/slip risk, and selects paths onboard.

    ControlConservative wheel bounding boxes, FPGA vision, and stop behavior.

  2. Step 2 of 2

    Earth operations team

    Sets goals and corridors, reviews telemetry, and handles hazards that stop autonomous progress.

    ControlHuman mission authority; ENav owns only bounded local motion.

Process model built from the published workflow evidence for NASA JPL. Every step, actor, and control appears in full below.
Every step, actor, and control

Exception path

Hazard or uncertainty stops rover for ground review.

Work removed

  • Human review between every short segment

Decision authority

ENav owns local motion; Earth owns mission goals.

Before

  1. 01

    Earth rover planners

    Use orbital and rover imagery to define route goals and commanded drive segments.

    Control: Communication delay prevents real-time joystick control; prior rover comparison provides no uniform terrain-adjusted cycle time.

  2. 02

    Rover and ground team

    Executes a commanded segment and waits for telemetry review before subsequent planning.

    Control: Ground mission safety review.

After

  1. 01

    Enhanced AutoNav

    Builds local terrain maps, checks collision/slip risk, and selects paths onboard.

    Control: Conservative wheel bounding boxes, FPGA vision, and stop behavior.

  2. 02

    Earth operations team

    Sets goals and corridors, reviews telemetry, and handles hazards that stop autonomous progress.

    Control: Human mission authority; ENav owns only bounded local motion.

Work that left the path

  • Human review between every short segment

Human role before

Earth-based rover planners translated imagery into route goals and commanded drive segments, then reviewed returned telemetry before queuing the next plan.

Human role after

Earth teams set goals, corridors and safety constraints.

AI roleEvaluates terrain, selects safe local paths and avoids obstacles onboard.

Outcomes

Autonomous distance share

Verified

Earlier rovers at most 5% in cited comparisonAbout 90% of 32.1 km used ENav

Through 2024-10-28 · 32.1 km

Cross-generation mission comparison.

What leaders can reuse

Anti-pattern

Do not compare speed without terrain context.

Questions

  1. 01Where is the operating threshold set and who can override it?
  2. 02What measured result would trigger rollback or retraining?
  3. 03Which residual decisions must remain human-owned?

Portability conditions

  • Bounded envelope
  • Conservative stop
  • Ground override

Reputation risk

low

Evidence and authority

What the public record supports.

Current · updated

1 peer reviewed, 1 primary; publication outcomes are verified.

Bundle 1.0.0 · reviewed 2026-08-23 · stable ID 08604fd5ea0768d6

Related transformations

More in Engineering design

Sources

Read the evidence, freshness, caveat, and version policy.