Update README.md

README.md

@@ -43,7 +43,7 @@ of the [ledrone](https://github.com/ahive-org/ledrone) project.
 
 | | |
 |---|---|
-| Task | `takeoff` — climb from the ground and hold a target altitude |
+| Task | `takeoff` : climb from the ground and hold a target altitude |
 | Type | Goal-conditioned, **state-based** (FPV image is an optional auxiliary) |
 | Episodes | 300 |
 | Frames | 60,000 (200/episode, 6.7s at 30 Hz) |
@@ -61,7 +61,7 @@ flies the `takeoff` reference in the `LeDrone` env on the **PyBullet** backend.
 
 - **DART-style coverage** ([Laskey et al. 2017](https://arxiv.org/abs/1703.09327)): the
   drone is driven by a *noised* command (σ = 0.15 m/s on the velocity channels) but each
-  frame is **labelled with the clean expert command** — so the policy sees
+  frame is **labelled with the clean expert command** : so the policy sees
   off-distribution states *and* the optimal correction for them, mitigating
   behavior-cloning covariate shift.
 - **Slew-rate-limited setpoint** ramps the command from rest, taming the low-level
@@ -89,7 +89,7 @@ python tools/generate_expert_dataset.py --task takeoff --backend pybullet \
 
 State is the SOTA quadrotor policy observation (relative goal + velocity + orientation +
 body rates); **battery is intentionally excluded** (handled by domain randomization, not
-observed — Hwangbo 2017, Molchanov 2019, Kaufmann ICRA 2022 / Nature 2023). Attitude is
+observed : Hwangbo 2017, Molchanov 2019, Kaufmann ICRA 2022 / Nature 2023). Attitude is
 a unit quaternion; map to a continuous 6D rep
 ([Zhou et al., CVPR 2019](https://arxiv.org/abs/1812.07035)) if desired. `action` is a
 NED velocity + yaw rate clipped to the envelope (≤ 2 m/s horizontal, ≤ 1 m/s climb,
@@ -121,15 +121,15 @@ frame = ds[0]  # observation.state (13,), observation.environment_state (3,), ac
 
 Analysis across all three `ledrone` tasks, computed from this DART-regenerated data.
 
-**Signature trajectories** — takeoff climbs from the ground, hover converges & station-keeps, circle flies full loops (>= 360 deg).
+**Signature trajectories** : takeoff climbs from the ground, hover converges & station-keeps, circle flies full loops (>= 360 deg).
 
 ![Trajectories](figures/trajectories.png)
 
-**Convergence & attitude** — every task drives the goal error to near-zero (circle keeps a small steady tracking lag from following a moving target); body tilt stays gentle (median ~10 deg, 95% <= 35 deg).
+**Convergence & attitude** : every task drives the goal error to near-zero (circle keeps a small steady tracking lag from following a moving target); body tilt stays gentle (median ~10 deg, 95% <= 35 deg).
 
 ![Convergence and attitude](figures/convergence_attitude.png)
 
-**Command / action space** — takeoff & hover station-keep at ~0.1 m/s while circle orbits at ~1.3 m/s; the climb command is bimodal for takeoff; yaw-rate is centered on the goal bearing.
+**Command / action space** : takeoff & hover station-keep at ~0.1 m/s while circle orbits at ~1.3 m/s; the climb command is bimodal for takeoff; yaw-rate is centered on the goal bearing.
 
 ![Action space](figures/action_space.png)